1 00:00:06,102 --> 00:00:10,707 >> HI, EVERYONE WELCOME TO THE 2 00:00:10,707 --> 00:00:12,108 IN SILICO DRUG DISCOVERY 3 00:00:12,108 --> 00:00:12,409 WORKSHOP. 4 00:00:12,409 --> 00:00:16,446 WE'RE EXCITED TO HAVE YOU JOIN 5 00:00:16,446 --> 00:00:17,747 US. 6 00:00:17,747 --> 00:00:23,086 I'M SER INE AT THE NATIONAL 7 00:00:23,086 --> 00:00:24,454 INSTITUTES OF HEALTH. 8 00:00:24,454 --> 00:00:27,056 SO WE HAVE A SUPER EXCITING 9 00:00:27,056 --> 00:00:29,426 PROGRAM FOR YOU ALL TODAY, WE'D 10 00:00:29,426 --> 00:00:33,930 LIKE TO START WITH OUR 11 00:00:33,930 --> 00:00:34,931 INTRODUCTORY COMMENTS. 12 00:00:34,931 --> 00:00:36,599 FOR THAT I'D LOVE TO INTRODUCE 13 00:00:36,599 --> 00:00:39,969 DR. JONI RUTTER WHO IS THE 14 00:00:39,969 --> 00:00:41,838 DIRECTOR OF THE NATIONAL CENTER 15 00:00:41,838 --> 00:00:44,574 FOR ADVANCING TRANSLATIONAL 16 00:00:44,574 --> 00:00:44,841 SCIENCES. 17 00:00:44,841 --> 00:00:45,942 JONI, WE'RE SUPER EXCITED TO 18 00:00:45,942 --> 00:00:48,878 HAVE YOU HERE TO OPEN UP THE 19 00:00:48,878 --> 00:00:49,145 WORKSHOP. 20 00:00:49,145 --> 00:00:51,481 THE FLOOR IS YOURS. 21 00:00:51,481 --> 00:00:53,516 >> THANK YOU, SO MUCH, SARINE. 22 00:00:53,516 --> 00:00:56,252 IN ADDITION TO BEING OUR EDITOR 23 00:00:56,252 --> 00:00:57,754 AND CHIEF OF THE ASSAY GUIDANCE 24 00:00:57,754 --> 00:01:04,494 MANUAL, SARINE IS AN AMAZING 25 00:01:04,494 --> 00:01:06,095 AMBASSADOR FOR TRANSLATIONAL 26 00:01:06,095 --> 00:01:08,965 SCIENCE AND THANK YOU TO OUR 27 00:01:08,965 --> 00:01:11,501 ORGANIZING TEAM AT NCATS, AND 28 00:01:11,501 --> 00:01:13,069 OUR COLLEAGUES AND I'M GLAD 29 00:01:13,069 --> 00:01:16,673 KIMBERLY WILL FOLLOW MY REMARKS 30 00:01:16,673 --> 00:01:16,973 TOO. 31 00:01:16,973 --> 00:01:24,247 AND I SHOUT OUT TO UC SAN DIEGO 32 00:01:24,247 --> 00:01:26,583 AND CHAPEL HILL AND IT'S SUPER 33 00:01:26,583 --> 00:01:26,850 EXCITING. 34 00:01:26,850 --> 00:01:31,521 AND THANK YOU ALL FOR JOINING US 35 00:01:31,521 --> 00:01:37,694 TODAY THIS IS THE OCTOBER AGM IN 36 00:01:37,694 --> 00:01:38,495 SILICO DRUG DISCOVERY WORKSHOP 37 00:01:38,495 --> 00:01:39,629 TODAY AND TOMORROW. 38 00:01:39,629 --> 00:01:40,897 IF YOU DON'T KNOW ABOUT THE 39 00:01:40,897 --> 00:01:43,566 PROGRAM IT'S REALLY A WORLD 40 00:01:43,566 --> 00:01:45,768 CLASS RESOURCE TO LOOK AT BEST 41 00:01:45,768 --> 00:01:48,538 PRACTICES IN TRANSLATIONAL 42 00:01:48,538 --> 00:01:51,241 SCIENCE AND IT PROVIDES ROBUST 43 00:01:51,241 --> 00:01:56,779 GUIDELINES FOR RESEARCH, ASSAY 44 00:01:56,779 --> 00:01:57,747 DEVELOPMENT, DATA ANALYSIS AND 45 00:01:57,747 --> 00:01:58,481 SO ON. 46 00:01:58,481 --> 00:02:00,884 WE HAVE DISCUSSIONS EQUALLY IN 47 00:02:00,884 --> 00:02:02,886 THESE AREAS OF SCIENCE AND 48 00:02:02,886 --> 00:02:06,322 ENSURE THE FINDINGS ARE 49 00:02:06,322 --> 00:02:07,957 REPRODUCIBLE AND THEY CAN MOVE 50 00:02:07,957 --> 00:02:11,628 SMOOTHLY FROM BENCH TO BEDSIDE. 51 00:02:11,628 --> 00:02:12,996 THESE WORKSHOPS PLAY A KEY ROLE 52 00:02:12,996 --> 00:02:14,297 IN HAVING THE CRITICAL 53 00:02:14,297 --> 00:02:14,998 DISCUSSIONS WITHIN THE 54 00:02:14,998 --> 00:02:15,865 SCIENTIFIC COMMUNITY TO MAKE 55 00:02:15,865 --> 00:02:17,734 THAT HAPPEN. 56 00:02:17,734 --> 00:02:22,138 THIS WORKSHOP WILL FOCUS ON 57 00:02:22,138 --> 00:02:23,806 INNOVATION WHICH IS ONE THE 58 00:02:23,806 --> 00:02:24,641 CRITICAL STEPS IN THE 59 00:02:24,641 --> 00:02:26,142 TRANSLATIONAL PIPELINE. 60 00:02:26,142 --> 00:02:29,279 I DON'T HAVE A SPECIAL EMPHASIS 61 00:02:29,279 --> 00:02:31,781 ON HOW DATA SCIENCE AND A.I. 62 00:02:31,781 --> 00:02:36,185 METHODS ARE REVOLUTIONIZING DRUG 63 00:02:36,185 --> 00:02:36,553 DISCOVERY. 64 00:02:36,553 --> 00:02:39,088 WE BELIEVE TEAM SCIENCE HAVING 65 00:02:39,088 --> 00:02:41,858 THESE TOGETHER WITH BIOLOGISTS, 66 00:02:41,858 --> 00:02:44,360 CHEMISTS, DATA SCIENTISTS AND 67 00:02:44,360 --> 00:02:45,128 STATISTICIANS ALL WORKING SIDE 68 00:02:45,128 --> 00:02:46,863 BY SIDE TO PUSH THE BOUNDARIES 69 00:02:46,863 --> 00:02:47,931 OF WHAT'S POSSIBLE. 70 00:02:47,931 --> 00:02:50,500 SO WE THINK IT'S CRITICAL TO 71 00:02:50,500 --> 00:02:53,736 HAVE THESE TO BRING ALL THOSE 72 00:02:53,736 --> 00:02:58,608 DISCIPLINES TOGETHER TO LOOK AT 73 00:02:58,608 --> 00:02:59,876 HOW WE CAN IMPLEMENT BEST 74 00:02:59,876 --> 00:03:00,176 PRACTICES. 75 00:03:00,176 --> 00:03:02,712 AS WE LOOK TO THE FUTURE, A.I. 76 00:03:02,712 --> 00:03:06,883 IS NOW PIVOTAL TO ACCELERATING 77 00:03:06,883 --> 00:03:09,886 DRUG DISCOVERY AND IMPROVING 78 00:03:09,886 --> 00:03:12,221 TRANSLATIONAL SCIENCE AND OUR 79 00:03:12,221 --> 00:03:15,391 TEAMS ENHANCE THINGS LIKE HIGH 80 00:03:15,391 --> 00:03:16,459 THROUGHPUT SCREENING AND 81 00:03:16,459 --> 00:03:18,895 AUTHORIZING COMPOUND PROPERTIES 82 00:03:18,895 --> 00:03:21,230 LIKE METABOLIC STABILITY FOR 83 00:03:21,230 --> 00:03:25,735 EXAMPLE AND THESE METHODS ALSO 84 00:03:25,735 --> 00:03:28,571 PREDICT SYNERGISTIC DRUG 85 00:03:28,571 --> 00:03:30,306 COMBINATIONS FOR CANCER AND RARE 86 00:03:30,306 --> 00:03:35,812 DISEASES AND ADVANCING THE NEXT 87 00:03:35,812 --> 00:03:40,583 GENERATION OF MODALITIES AND 88 00:03:40,583 --> 00:03:42,885 TARGETED ANTIBODIES THESE ARE 89 00:03:42,885 --> 00:03:45,755 ALL CRITICAL MODALITIES COMING 90 00:03:45,755 --> 00:03:47,457 ON SCENE AND HOW WE STUDY THEM 91 00:03:47,457 --> 00:03:53,363 AND HOW WE GENERATE THEM AND HOW 92 00:03:53,363 --> 00:03:55,765 WE TRANSLATING THEM TO CLINICAL 93 00:03:55,765 --> 00:03:57,567 PRACTICE IS WHERE THIS IS ALL 94 00:03:57,567 --> 00:03:57,834 STARTING. 95 00:03:57,834 --> 00:04:02,338 ONE WAY NCATS APPLIES A.I. IN 96 00:04:02,338 --> 00:04:03,640 IMPROVING HIGH THROUGHPUT 97 00:04:03,640 --> 00:04:06,643 SCREENING OUTCOMES BY DETECTING 98 00:04:06,643 --> 00:04:07,944 POTENTIAL ASSAY ARTIFACTS 99 00:04:07,944 --> 00:04:09,746 ENSURING THE DATA WE RELY ON IS 100 00:04:09,746 --> 00:04:11,247 ROBUST AND ACCURATE. 101 00:04:11,247 --> 00:04:12,081 THAT'S CRITICAL FOR ADVANCING 102 00:04:12,081 --> 00:04:15,418 THE PROMISING CANDIDATES 103 00:04:15,418 --> 00:04:16,085 FORWARD. 104 00:04:16,085 --> 00:04:18,821 WE HAVE A REALLY HIGH FAILURE 105 00:04:18,821 --> 00:04:20,390 RATE WHEN WE MOVE COMPOUNDS 106 00:04:20,390 --> 00:04:21,758 THROUGH THE TRANSLATIONAL 107 00:04:21,758 --> 00:04:22,058 PIPELINE. 108 00:04:22,058 --> 00:04:24,927 THIS IS AN AREA WE HOPE TO 109 00:04:24,927 --> 00:04:26,329 ADDRESS TO BRING DOWN THAT 110 00:04:26,329 --> 00:04:27,163 FAILURE RATE. 111 00:04:27,163 --> 00:04:28,564 WE'RE ALSO THINKING ABOUT 112 00:04:28,564 --> 00:04:31,401 APPLICATIONS IN QUANTUM SCIENCES 113 00:04:31,401 --> 00:04:32,735 AND HAVE CHALLENGE COMPETITIONS 114 00:04:32,735 --> 00:04:35,171 OUT ON THE STREET ON THAT. 115 00:04:35,171 --> 00:04:36,739 I'M HAPPY TO PUT THOSE IN THE 116 00:04:36,739 --> 00:04:39,742 LINKS WHEN I'M FINISHED HERE. 117 00:04:39,742 --> 00:04:41,778 BUT THE THREE AREAS OF INTEREST 118 00:04:41,778 --> 00:04:44,414 INCLUDE QUANTUM ALGORITHMS FOR 119 00:04:44,414 --> 00:04:47,183 DRUG DISCOVERY AND FOR CLINICAL 120 00:04:47,183 --> 00:04:52,855 RISK PREDICTIONS, DIAGNOSIS AND 121 00:04:52,855 --> 00:04:56,793 THERAPEUTICS AS WELL AS QUANTUM 122 00:04:56,793 --> 00:05:00,530 ALGORITHMS AND TO IDENTIFY GAPS 123 00:05:00,530 --> 00:05:01,531 IN THE TRANSLATION OF IN SILICO 124 00:05:01,531 --> 00:05:03,766 MODELS TO EFFECTIVE THERAPIES. 125 00:05:03,766 --> 00:05:05,735 AS WE'VE SEEN CHALLENGING IN 126 00:05:05,735 --> 00:05:07,603 TRANSLATING DATA FROM BIO 127 00:05:07,603 --> 00:05:09,839 CHEMICAL ASSAYS TO CELL-BASED 128 00:05:09,839 --> 00:05:11,007 ASSAYS AND TISSUES AND ANIMALS 129 00:05:11,007 --> 00:05:15,845 TO HUMANS, SIMILAR GAPS EXIST IN 130 00:05:15,845 --> 00:05:18,481 THE IN SILICO MODELS TO THEN THE 131 00:05:18,481 --> 00:05:20,283 REAL WORLD APPLICATIONS THAT 132 00:05:20,283 --> 00:05:21,517 LEAD TO EFFECTIVE TREATMENTS. 133 00:05:21,517 --> 00:05:27,023 EACH STAGE OF THE DRUG DISCOVERY 134 00:05:27,023 --> 00:05:27,890 PROCESS REQUIRES THOSE SPECIFIC 135 00:05:27,890 --> 00:05:30,393 TYPES OF MODELS AND THE 136 00:05:30,393 --> 00:05:31,661 PREDICTIONS MADE BY THESE MODELS 137 00:05:31,661 --> 00:05:37,734 HAS TO ALIGN WITH THE BIOLOGICAL 138 00:05:37,734 --> 00:05:38,067 COMPLEXITIES. 139 00:05:38,067 --> 00:05:43,239 AS WE MOVE FROM COMPUTATIONAL 140 00:05:43,239 --> 00:05:44,507 PREDICTS WE NEED TO LOOK AT WHEN 141 00:05:44,507 --> 00:05:46,943 MODELS AND PREDICTIONS ARE BEST 142 00:05:46,943 --> 00:05:48,344 SUITED FOR EACH STEP IN THE 143 00:05:48,344 --> 00:05:48,644 PIPELINE. 144 00:05:48,644 --> 00:05:50,480 WE ALSO NEED TO CONSIDER WHAT 145 00:05:50,480 --> 00:05:52,181 NEW MODELS AND APPROACHES ARE 146 00:05:52,181 --> 00:05:54,217 REQUIRES TO IMPROVE TRANSLATION 147 00:05:54,217 --> 00:05:55,885 AND ENSURE THE IN SILICO 148 00:05:55,885 --> 00:05:57,220 INSIGHTS LEAD TO THOSE TANGIBLE 149 00:05:57,220 --> 00:06:02,625 THERAPEUTIC ADVANCEMENTS. 150 00:06:02,625 --> 00:06:03,359 SO, THAT'S WHAT I'M SUPER 151 00:06:03,359 --> 00:06:05,061 EXCITED ABOUT FOR THE MEETING IS 152 00:06:05,061 --> 00:06:10,900 GETTING GOOD INSIGHT TO TALK 153 00:06:10,900 --> 00:06:11,634 ABOUT THIS. 154 00:06:11,634 --> 00:06:13,736 AND I WANT TO ACKNOWLEDGE 155 00:06:13,736 --> 00:06:15,138 SOMETHING TRULY EXTRAORDINARY. 156 00:06:15,138 --> 00:06:16,973 IT'S NOT EVERY DAY WE GET TO 157 00:06:16,973 --> 00:06:20,710 HEAR FROM A NEWLY-MINTED NOBEL 158 00:06:20,710 --> 00:06:24,747 LAUREATE AND GLAD TO HAVE DAVID 159 00:06:24,747 --> 00:06:24,981 BEAKER. 160 00:06:24,981 --> 00:06:27,450 HE WAS AWARDED THE 2024 NOBEL 161 00:06:27,450 --> 00:06:29,318 PRIZE IN CHEMISTRY FOR HIS 162 00:06:29,318 --> 00:06:30,887 GROUNDBREAKING WORK IN 163 00:06:30,887 --> 00:06:33,756 COMPUTATIONAL PROTEIN DESIGN 164 00:06:33,756 --> 00:06:36,526 WITH DENISE AND JOHN WITH THEIR 165 00:06:36,526 --> 00:06:37,760 PROTEIN STRUCTURE WORK TOO. 166 00:06:37,760 --> 00:06:40,163 THIS FIELD IS FUNDAMENTALLY 167 00:06:40,163 --> 00:06:41,264 RESHAPING OUR UNDERSTANDING OF 168 00:06:41,264 --> 00:06:44,867 BIOLOGY AND OPENING NEW DOORS 169 00:06:44,867 --> 00:06:46,135 FOR THERAPEUTIC DEVELOPMENT AND 170 00:06:46,135 --> 00:06:48,538 LOOKING FORWARD TO THESE TALKS 171 00:06:48,538 --> 00:06:50,306 AS WE START TO HEAR MORE AND 172 00:06:50,306 --> 00:06:51,974 APPLY THOSE LEARN TO WHAT WE'RE 173 00:06:51,974 --> 00:07:02,518 GOING TO DO IN THE COMING YEARS. 174 00:07:07,023 --> 00:07:08,691 WE'LL LOOK AT THE PRACTICAL 175 00:07:08,691 --> 00:07:10,126 CHALLENGES OF TRANSLATING 176 00:07:10,126 --> 00:07:11,160 ANTIRETROVIRAL MODELS INTO 177 00:07:11,160 --> 00:07:12,128 EFFECTIVE THERAPIES. 178 00:07:12,128 --> 00:07:13,763 THIS IS IMPORTANT TO NOTE 179 00:07:13,763 --> 00:07:16,999 BECAUSE THE NOBEL PRIZE IN 180 00:07:16,999 --> 00:07:19,769 PHYSICS WAS AWARDED FOR 181 00:07:19,769 --> 00:07:21,737 FOUNDATIONAL DISCOVERIES AND 182 00:07:21,737 --> 00:07:23,606 INVENTIONS THAT ENABLE MACHINE 183 00:07:23,606 --> 00:07:26,375 LEARNING WITH ARTIFICIAL NEURAL 184 00:07:26,375 --> 00:07:26,642 NETWORKS. 185 00:07:26,642 --> 00:07:28,177 WHILE IT'S REALLY FANTASTIC TO 186 00:07:28,177 --> 00:07:29,879 CELEBRATE THOSE ACHIEVEMENTS, 187 00:07:29,879 --> 00:07:32,148 IT'S ALSO EQUALLY IMPORTANT THAT 188 00:07:32,148 --> 00:07:33,749 WE BUILD ON THEM. 189 00:07:33,749 --> 00:07:35,084 IDENTIFYING GAPS IN THE 190 00:07:35,084 --> 00:07:36,886 TRANSLATIONAL PROCESS AND 191 00:07:36,886 --> 00:07:41,824 FINDING WAYS TO ADDRESS THEM. 192 00:07:41,824 --> 00:07:45,027 HOW DO WE ENSURE WHAT WORKS IN 193 00:07:45,027 --> 00:07:46,729 THE LABS WORKS IN REAL WORLD 194 00:07:46,729 --> 00:07:48,030 SETTING AND IMPROVE THE 195 00:07:48,030 --> 00:07:49,732 RELIABILITY OF DRUG DEVELOPMENT? 196 00:07:49,732 --> 00:07:51,300 THESE ARE KEY QUESTIONS I LOOK 197 00:07:51,300 --> 00:07:53,236 FORWARD TO LARGER ABOUT OVER THE 198 00:07:53,236 --> 00:07:53,736 NEXT TWO DAYS FOR THIS 199 00:07:53,736 --> 00:07:55,738 CONFERENCE. 200 00:07:55,738 --> 00:07:57,406 AND THIS WORKSHOP REALLY 201 00:07:57,406 --> 00:07:58,741 REPRESENTS THE COLLECTIVE 202 00:07:58,741 --> 00:07:59,976 EFFORTS OF INSTITUTIONS ACROSS 203 00:07:59,976 --> 00:08:01,744 THE PUBLIC AND PRIVATE SECTORS 204 00:08:01,744 --> 00:08:07,516 THAT ARE EXEMPLIFYING THE POWER 205 00:08:07,516 --> 00:08:09,752 AND COLLABORATION OF TEAM 206 00:08:09,752 --> 00:08:11,587 SCIENCE TO DRIVE THE SCIENCE 207 00:08:11,587 --> 00:08:11,888 FORWARD. 208 00:08:11,888 --> 00:08:13,756 AS WE KICK OFF THE MEETING TODAY 209 00:08:13,756 --> 00:08:15,224 ONLY THING I ASKED FOR YOU TO 210 00:08:15,224 --> 00:08:16,292 KEEP THINKING ABOUT IN THE BACK 211 00:08:16,292 --> 00:08:17,560 OF YOUR MIND THAT MOVES TO THE 212 00:08:17,560 --> 00:08:20,897 FRONT OF THE YOUR MIND IS PLEASE 213 00:08:20,897 --> 00:08:22,999 DON'T LOSE SIGHT OF THE REAL 214 00:08:22,999 --> 00:08:23,633 GOAL HERE. 215 00:08:23,633 --> 00:08:24,934 THE WORK WE DO HERE TODAY WILL 216 00:08:24,934 --> 00:08:28,804 SHAPE OU QUICKLY AND EFFECTIVELY 217 00:08:28,804 --> 00:08:30,873 NEW TREATMENTS CAN REACH 218 00:08:30,873 --> 00:08:33,276 PATIENTS AND SAVE LIVES. 219 00:08:33,276 --> 00:08:35,044 WE LIKE TO THINK TO BRING NEW 220 00:08:35,044 --> 00:08:36,445 TREATMENTS FOR ALL PEOPLE MORE 221 00:08:36,445 --> 00:08:36,913 QUICKLY. 222 00:08:36,913 --> 00:08:40,516 THANKS FOR JOINING TODAY AND 223 00:08:40,516 --> 00:08:42,885 TOMORROW AND LET'S GET STARTED. 224 00:08:42,885 --> 00:08:47,590 KIMBERLY, OVER TO YOU OR SARINE. 225 00:08:47,590 --> 00:08:47,857 THANK YOU. 226 00:08:47,857 --> 00:08:55,164 >> THANK YOU SO MUCH, JOANIE. 227 00:08:55,164 --> 00:08:57,733 SO, THE NEXT PERSON FOR THE 228 00:08:57,733 --> 00:09:01,170 OPENING REMARKS IS DR. KIMBERLY 229 00:09:01,170 --> 00:09:01,470 SCIARRETTA. 230 00:09:01,470 --> 00:09:03,472 SHE'S THE ACTING DIRECTOR 231 00:09:03,472 --> 00:09:07,843 DIVISION OF RESEARCH AND 232 00:09:07,843 --> 00:09:12,915 VENTURES DRIVE PART OF THE 233 00:09:12,915 --> 00:09:14,116 BIOMEDICAL ADVANCED RESEARCH AND 234 00:09:14,116 --> 00:09:16,385 DEVELOPMENT AUTHORITY, BARDA. 235 00:09:16,385 --> 00:09:18,921 THANK YOU FOR JOINING US. 236 00:09:18,921 --> 00:09:29,265 THE FLOOR IS YOURS. 237 00:09:40,476 --> 00:09:43,579 >> THANK YOU, SARINE. 238 00:09:43,579 --> 00:09:45,581 IT DOESN'T LOOK LIKE I CAN BE ON 239 00:09:45,581 --> 00:09:47,750 VIDEO SO MAYBE THAT CAN BE FIXED 240 00:09:47,750 --> 00:09:53,723 AND HOPE THE LARYNGITIS DIDN'T 241 00:09:53,723 --> 00:09:56,492 STOP ME AND I'M PLEASED TO KICK 242 00:09:56,492 --> 00:09:57,393 OFF THE MEETING WITH OUR 243 00:09:57,393 --> 00:10:00,062 COLLEAGUES AT NCATS. 244 00:10:00,062 --> 00:10:02,598 I HEARD WE'VE HAD QUITE OF 245 00:10:02,598 --> 00:10:04,533 RESPONSE AND REFLECTION OF HOW 246 00:10:04,533 --> 00:10:05,735 RESONATING AND IMPACTFUL THE 247 00:10:05,735 --> 00:10:06,902 TOPIC IS TO ALL OF YOU. 248 00:10:06,902 --> 00:10:11,574 I WOULD BE REMISS IF WE DIDN'T 249 00:10:11,574 --> 00:10:14,276 THANK AND ACKNOWLEDGE ALL THE 250 00:10:14,276 --> 00:10:15,611 WORKSHOP ORGANIZING COMMITTEE 251 00:10:15,611 --> 00:10:17,346 AND EFFORTS PUT INTO PLANNING 252 00:10:17,346 --> 00:10:18,914 THE MEETING WITH OUR COLLEAGUES 253 00:10:18,914 --> 00:10:24,620 AT NCATS, BARDA AND UCSD AND 254 00:10:24,620 --> 00:10:24,920 UNC. 255 00:10:24,920 --> 00:10:26,789 BEFORE I DISCUSS OUR INTEREST IN 256 00:10:26,789 --> 00:10:29,759 THIS SPACING AND WHY WE'RE 257 00:10:29,759 --> 00:10:31,794 EXCITED ABOUT THE MEETING AND 258 00:10:31,794 --> 00:10:34,997 WANTED TO PROVIDE BACKGROUND ON 259 00:10:34,997 --> 00:10:35,698 BARDA. 260 00:10:35,698 --> 00:10:37,466 THE BIOMEDICAL ADVANCED RESEARCH 261 00:10:37,466 --> 00:10:38,501 AND DEVELOPMENT AUTHORITY WITHIN 262 00:10:38,501 --> 00:10:41,737 THE ADMINISTRATION OF STRATEGIC 263 00:10:41,737 --> 00:10:44,073 PREPAREDNESS AND RESPONSE WHICH 264 00:10:44,073 --> 00:10:45,741 IS WITHIN THE HHS. 265 00:10:45,741 --> 00:10:48,911 I KNOW QUITE A MOUTHFUL. 266 00:10:48,911 --> 00:10:51,313 THE MISSION IS TO PREPARE AND 267 00:10:51,313 --> 00:10:51,914 RESPOND TO PUBLIC HEALTH 268 00:10:51,914 --> 00:10:53,282 EMERGENCIES AND A BIG PART OF 269 00:10:53,282 --> 00:10:54,917 OUR WORK WAS RESPONDING TO 270 00:10:54,917 --> 00:10:57,753 COVID-19 AND WE RESPOND TO OTHER 271 00:10:57,753 --> 00:10:59,488 PUBLIC HEALTH DISASTERS LIKE 272 00:10:59,488 --> 00:11:02,425 HURRICANES AND FOREST FIRES. 273 00:11:02,425 --> 00:11:04,260 BARDA'S ROLE IS TO DEVELOP THE 274 00:11:04,260 --> 00:11:05,494 COUNTER MEASURES AND MAKE THEM 275 00:11:05,494 --> 00:11:07,763 AVAILABLE FOR THE PUBLIC HEALTH 276 00:11:07,763 --> 00:11:08,264 EMERGENCIES. 277 00:11:08,264 --> 00:11:14,437 WHAT DO I MEAN BY MEDICAL CO 278 00:11:14,437 --> 00:11:16,205 COUNTER MEASURES AND 279 00:11:16,205 --> 00:11:17,139 DIAGNOSTICS, VACCINES, ANY TYPE 280 00:11:17,139 --> 00:11:18,074 OF THERAPEUTICS AND DRUGS WHICH 281 00:11:18,074 --> 00:11:22,645 IS THE TOPIC OF THIS MEETING. 282 00:11:22,645 --> 00:11:23,979 AND WE'RE DEVELOPING THESE 283 00:11:23,979 --> 00:11:28,717 MEDICAL COUNTER MEASURES TO 284 00:11:28,717 --> 00:11:33,422 RESPOND TO ANY OF THESE FLU AND 285 00:11:33,422 --> 00:11:36,926 EMERGING INFECTIOUS DISEASES 286 00:11:36,926 --> 00:11:39,361 WITH UNIQUE PUBLIC-PRIVATE 287 00:11:39,361 --> 00:11:41,630 PARTNERSHIP AND OUR PARTNERS 288 00:11:41,630 --> 00:11:43,766 HAVE HAD A TOTAL OF 95 FDA 289 00:11:43,766 --> 00:11:45,701 APPROVALS AND LICENSES AND 290 00:11:45,701 --> 00:11:47,269 CLEARANCES TO DEVELOPMENT OF 291 00:11:47,269 --> 00:11:51,340 COUNTER MEASURES THAT WILL 292 00:11:51,340 --> 00:11:51,907 DIRECTLY PROTECT AMERICANS 293 00:11:51,907 --> 00:11:52,508 FACING PUBLIC HEALTH 294 00:11:52,508 --> 00:11:52,808 EMERGENCIES. 295 00:11:52,808 --> 00:11:54,243 THIS IS A TESTAMENT TO THE 296 00:11:54,243 --> 00:11:57,746 PARTNERSHIPS OF WORKING WITH 297 00:11:57,746 --> 00:12:00,916 ORGANIZATIONS LIKE YOURSELVES. 298 00:12:00,916 --> 00:12:02,451 SO BARDA TRADITIONALLY FOCUSES 299 00:12:02,451 --> 00:12:04,353 ON ADVANCED RESEARCH AND 300 00:12:04,353 --> 00:12:06,889 DEVELOPMENT OF MEDICAL COUNTER 301 00:12:06,889 --> 00:12:07,156 MEASURES. 302 00:12:07,156 --> 00:12:09,091 HOWEVER WE HAVE ESTABLISHED THE 303 00:12:09,091 --> 00:12:11,026 DRIVE DIVISION WHICH PLAYS MORE 304 00:12:11,026 --> 00:12:12,528 IN THIS INNOVATION SPACE 305 00:12:12,528 --> 00:12:14,864 SUPPORTING ENABLING TECHNOLOGIES 306 00:12:14,864 --> 00:12:16,232 AND SUPPORTING SOME EARLIER 307 00:12:16,232 --> 00:12:16,899 RESEARCH ESPECIALLY THROUGH SOME 308 00:12:16,899 --> 00:12:21,203 OF OUR PARTNERSHIPS WITH BLUE 309 00:12:21,203 --> 00:12:27,343 NIGHT AND THE BARDA ACCELERATOR 310 00:12:27,343 --> 00:12:37,887 NETWORK AND OUR VENTURES TO ECHO 311 00:12:39,488 --> 00:12:40,756 DR. RUTTER WE'RE IN A UNIQUE 312 00:12:40,756 --> 00:12:44,693 POSITION TO BE KICKING OFF THE 313 00:12:44,693 --> 00:12:48,197 MEETING AT A TIME WHEN NOBEL 314 00:12:48,197 --> 00:12:49,932 PRIZES WERE ANNOUNCED FOR 315 00:12:49,932 --> 00:12:52,501 CHEMISTRY AND PHYSICS FOR THIS 316 00:12:52,501 --> 00:12:52,902 TOPIC. 317 00:12:52,902 --> 00:12:56,939 THOUGH WE ALL UNDERSTAND THE 318 00:12:56,939 --> 00:12:57,740 POTENTIAL VALUE IN SILICO BRINGS 319 00:12:57,740 --> 00:13:00,910 TO DRUG DISCOVERY AND MANY HAVE 320 00:13:00,910 --> 00:13:02,311 EXPERIENCE THE LENGTHY TIME AND 321 00:13:02,311 --> 00:13:03,445 IDENTIFICATION OF THESE 322 00:13:03,445 --> 00:13:07,516 CANDIDATES, WE DO SEE THE VALUE 323 00:13:07,516 --> 00:13:09,218 THESE APPROACHES CAN REALLY 324 00:13:09,218 --> 00:13:11,453 BRING TO IDENTIFYING SAFER AND 325 00:13:11,453 --> 00:13:12,922 MORE EFFICACIOUS DRUGS FASTER 326 00:13:12,922 --> 00:13:14,456 AND SOONER AND THIS IS 327 00:13:14,456 --> 00:13:16,659 CRITICALLY IMPORTANT FOR OUR 328 00:13:16,659 --> 00:13:20,162 ABILITY TO RESPOND TO PUBLIC 329 00:13:20,162 --> 00:13:21,063 HEALTH EMERGENCIES. 330 00:13:21,063 --> 00:13:23,599 I DID WANT TO ACKNOWLEDGE AGAIN 331 00:13:23,599 --> 00:13:27,536 THE GREAT WORK BY OUR COLLEAGUES 332 00:13:27,536 --> 00:13:29,772 IN THIS SPACE AND HOW IMPACTFUL 333 00:13:29,772 --> 00:13:36,912 THIS WORK REALLY IS AND WE LOOK 334 00:13:36,912 --> 00:13:38,914 FORWARD TO HEARING FROM 335 00:13:38,914 --> 00:13:41,283 DR. BAKER AND OUR ESTEEMED 336 00:13:41,283 --> 00:13:41,750 COLLEAGUES. 337 00:13:41,750 --> 00:13:43,619 THE WHOLE MEETING WOULD NOT BE A 338 00:13:43,619 --> 00:13:45,688 SUCCESS IF IT WASN'T FOR ALL OF 339 00:13:45,688 --> 00:13:45,854 YOU. 340 00:13:45,854 --> 00:13:46,822 WE'RE BRINGING TOGETHER 341 00:13:46,822 --> 00:13:49,325 U.S. GOVERNMENT WITH ACADEMIC 342 00:13:49,325 --> 00:13:51,994 AND INDUSTRY EXPERTS TO EXPLORE 343 00:13:51,994 --> 00:13:55,698 THE FUTURE OF IN SILICO 344 00:13:55,698 --> 00:13:55,965 MODELLING. 345 00:13:55,965 --> 00:14:01,704 SO AN ASK IS TO FOSTER AND FORM 346 00:14:01,704 --> 00:14:02,738 COLLABORATIONS AS MUCH AS 347 00:14:02,738 --> 00:14:03,739 POSSIBLE IN THIS VIRTUAL 348 00:14:03,739 --> 00:14:04,173 ENVIRONMENT. 349 00:14:04,173 --> 00:14:06,775 WE WANT YOU TO ENGAGE IN 350 00:14:06,775 --> 00:14:07,309 INFORMATION SHARING AND 351 00:14:07,309 --> 00:14:09,011 HOPEFULLY GENERATING NEW IDEAS 352 00:14:09,011 --> 00:14:10,646 TO ADDRESS CHALLENGES. 353 00:14:10,646 --> 00:14:12,514 I HOPE WE CAN BRAIN STORM 354 00:14:12,514 --> 00:14:13,148 SOLUTIONS TOGETHER AND MAYBE 355 00:14:13,148 --> 00:14:14,817 SOME WILL BE THE GENESIS OF NEW 356 00:14:14,817 --> 00:14:16,218 PROGRAM IDEAS. 357 00:14:16,218 --> 00:14:17,753 HOPEFULLY WE'LL ALL GAIN A 358 00:14:17,753 --> 00:14:19,622 BETTER UNDERSTANDING OF THE 359 00:14:19,622 --> 00:14:21,757 CURRENT LANDSCAPE OF 360 00:14:21,757 --> 00:14:23,492 COMPUTATIONAL DRUG DISCOVERY. 361 00:14:23,492 --> 00:14:27,229 AGAIN, THIS INFORMATION WILL 362 00:14:27,229 --> 00:14:28,897 DIRECTLY HELP US PREPARE FOR 363 00:14:28,897 --> 00:14:29,765 FUTURE PUBLIC HEALTH 364 00:14:29,765 --> 00:14:39,975 EMERGENCIES. 365 00:14:42,077 --> 00:14:44,446 BARDA CURRENTLY DOESN'T HAVE 366 00:14:44,446 --> 00:14:45,581 OPENING AND MOST WORK IS 367 00:14:45,581 --> 00:14:46,548 FOCUSSED ON ADVANCED RESEARCH 368 00:14:46,548 --> 00:14:49,718 AND DEVELOPMENT AND THESE 369 00:14:49,718 --> 00:14:51,053 ACTIVITIES TYPICALLY OCCUR MORE 370 00:14:51,053 --> 00:14:52,688 UP STREAM WITH OUR PARTNERS 371 00:14:52,688 --> 00:14:55,391 BEFORE THEY ENGAGE WITH BARDA. 372 00:14:55,391 --> 00:14:57,059 HOWEVER, AS I NOTED WE'RE 373 00:14:57,059 --> 00:14:58,594 INTERESTED IN INNOVATIVE 374 00:14:58,594 --> 00:14:59,628 OPPORTUNITIES AND ENABLING 375 00:14:59,628 --> 00:15:01,730 TECHNOLOGIES ESPECIALLY WITHIN 376 00:15:01,730 --> 00:15:04,700 THE DRUG DIVISION AND THERE WAS 377 00:15:04,700 --> 00:15:06,402 A NOTABLE EFFORT IN OUR PROGRAM 378 00:15:06,402 --> 00:15:06,902 REDIRECT. 379 00:15:06,902 --> 00:15:10,406 IT'S A PARTNERSHIP WITH DRIVE 380 00:15:10,406 --> 00:15:12,908 AND OUR DIVISION CBRN FOCUSSED 381 00:15:12,908 --> 00:15:15,010 ON REPURPOSING DRUGS TO TREAT 382 00:15:15,010 --> 00:15:16,512 CONDITIONS AS A RESPONSE TO 383 00:15:16,512 --> 00:15:19,348 EXPOSURE TO CHEMICAL AGENTS AND 384 00:15:19,348 --> 00:15:20,716 THERE WERE EFFORTS IN THIS 385 00:15:20,716 --> 00:15:23,719 PROGRAM RELATED TO IN SILICO 386 00:15:23,719 --> 00:15:24,019 APPROACHES. 387 00:15:24,019 --> 00:15:24,787 AGAIN, THERE'S OPPORTUNITY AS WE 388 00:15:24,787 --> 00:15:25,854 LEARNED FROM THE MEETING TO 389 00:15:25,854 --> 00:15:29,391 CONTINUE TO EXPAND IN THIS 390 00:15:29,391 --> 00:15:29,591 SPACE. 391 00:15:29,591 --> 00:15:30,926 I DID WANT TO HIGHLIGHT THE 392 00:15:30,926 --> 00:15:32,227 VALUE AND TRANSLATIONAL SCIENCE 393 00:15:32,227 --> 00:15:34,363 HERE OF THIS WORK WE'RE GOING TO 394 00:15:34,363 --> 00:15:36,365 HEAR TODAY AND HOW THAT'S GO TO 395 00:15:36,365 --> 00:15:37,766 APPLY TO FUTURE EVENTS AND 396 00:15:37,766 --> 00:15:38,133 DEVELOPMENT. 397 00:15:38,133 --> 00:15:41,136 THIS WILL DIRECTLY IMPACT ALL OF 398 00:15:41,136 --> 00:15:44,106 YOU AND HOPEFULLY SAVE AMERICANS 399 00:15:44,106 --> 00:15:45,741 WHEN WE ARE FACING A PUBLIC 400 00:15:45,741 --> 00:15:48,010 HEALTH THREAT IN THE FUTURE. 401 00:15:48,010 --> 00:15:52,247 SO ONE RAPID DRUG IDENTIFICATION 402 00:15:52,247 --> 00:15:53,349 IMMENSERY CRITICAL WHEN WE ARE 403 00:15:53,349 --> 00:15:56,118 FACING A PUBLIC HEALTH EMERGENCY 404 00:15:56,118 --> 00:15:59,788 AND THIS COULD BE IN THE DESIGN 405 00:15:59,788 --> 00:16:01,757 OF NOVEL DRUGS AND CANDIDATE TO 406 00:16:01,757 --> 00:16:02,358 CLINICALLY EVALUATE. 407 00:16:02,358 --> 00:16:03,959 THE SECOND IS WE NEED TO 408 00:16:03,959 --> 00:16:06,795 INCREASE OUR ABILITY TO APPLY 409 00:16:06,795 --> 00:16:09,498 OUR LEARN FROM PREVIOUS THREATS 410 00:16:09,498 --> 00:16:11,500 WE'VE SEEN TO NEW EMERGING 411 00:16:11,500 --> 00:16:11,800 THREATS. 412 00:16:11,800 --> 00:16:17,740 THAT TOUCHES ON OUR RESPONSE 413 00:16:17,740 --> 00:16:18,741 CAPABILITIES AND BUILDING OUT 414 00:16:18,741 --> 00:16:21,143 PLATFORMS AND MODELS AND 415 00:16:21,143 --> 00:16:22,644 APPROACHES WILL ALLOW US TO BE 416 00:16:22,644 --> 00:16:24,413 MORE AGILE IN RESPONSE TO FUTURE 417 00:16:24,413 --> 00:16:24,880 THREATS. 418 00:16:24,880 --> 00:16:27,983 THE OTHER UNIQUE THING ABOUT 419 00:16:27,983 --> 00:16:29,785 DEVELOPING MEDICAL COUNTER 420 00:16:29,785 --> 00:16:33,389 MEASURES FOR CBRN THREATS IS WE 421 00:16:33,389 --> 00:16:34,957 CAN'T ALWAYS EVALUATE THEM IN 422 00:16:34,957 --> 00:16:35,991 CLINICAL MODELS. 423 00:16:35,991 --> 00:16:37,326 THEREFORE THERE'S A NEED FOR 424 00:16:37,326 --> 00:16:39,128 SUFFICIENT AND EQUIVALENT MODELS 425 00:16:39,128 --> 00:16:40,996 TO EVALUATE THESE PRODUCTS EVEN 426 00:16:40,996 --> 00:16:42,431 OUTSIDE OF ANIMALS. 427 00:16:42,431 --> 00:16:43,866 OF COURSE WE'RE ALWAYS 428 00:16:43,866 --> 00:16:45,734 INTERESTED IN UNDERSTANDING 429 00:16:45,734 --> 00:16:47,035 PHARMACOLOGY AND TOXICOLOGY OF 430 00:16:47,035 --> 00:16:48,003 THE DRUGS. 431 00:16:48,003 --> 00:16:50,472 THESE ARE SEVERAL AREAS WHERE WE 432 00:16:50,472 --> 00:16:53,742 SEE THE VALUE OF IN SILICO 433 00:16:53,742 --> 00:16:55,310 APPROACHING BARDA AND THE DRUGS 434 00:16:55,310 --> 00:16:59,515 THAT CAN GET TO YOU ALL SOONER. 435 00:16:59,515 --> 00:17:00,983 THERE ARE SEVERAL OFFICES WITHIN 436 00:17:00,983 --> 00:17:03,585 BARDA FOCUSSED ON MEDICAL 437 00:17:03,585 --> 00:17:07,322 COUNTER MEASURE DEVELOPMENT. 438 00:17:07,322 --> 00:17:11,660 CBRN, FLU, DDDI AND DRIVE. 439 00:17:11,660 --> 00:17:13,629 THAT'S THE DIVISION I SIT. 440 00:17:13,629 --> 00:17:15,330 OUR MICKS IS TO SUPPORT THE 441 00:17:15,330 --> 00:17:17,733 DEVELOPMENT AND ADOPTION OF 442 00:17:17,733 --> 00:17:19,301 THESE INNOVATIVE TECHNOLOGIES 443 00:17:19,301 --> 00:17:21,403 AND COUNTER MEASURES. 444 00:17:21,403 --> 00:17:23,005 WE DON'T JUST SUPPORT THE 445 00:17:23,005 --> 00:17:24,239 COUNTER MEASURE DEVELOPMENT WE 446 00:17:24,239 --> 00:17:25,374 SUPPORT THE ENABLING 447 00:17:25,374 --> 00:17:26,708 TECHNOLOGIES THAT ALLOW THE 448 00:17:26,708 --> 00:17:28,444 PRODUCTS TO BE SUCCESSFUL. 449 00:17:28,444 --> 00:17:30,913 YOU CAN ENGAGE DIRECTLY WITH US 450 00:17:30,913 --> 00:17:32,548 IF YOU'RE INTERESTED IN REACHING 451 00:17:32,548 --> 00:17:34,383 OUT AT HHS.gov YOU CAN ALSO 452 00:17:34,383 --> 00:17:36,752 CHECK OUT SOME OF OUR PROGRAMS 453 00:17:36,752 --> 00:17:39,588 AT DRIVE.HHS.gov. 454 00:17:39,588 --> 00:17:43,192 YOU CAN ALSO ENGAGE WITH THE 455 00:17:43,192 --> 00:17:47,296 REST OF BARDA IF YOU THERES OF 456 00:17:47,296 --> 00:17:47,563 INTEREST. 457 00:17:47,563 --> 00:17:48,464 EVERY SOLICITATION TOPIC HAS 458 00:17:48,464 --> 00:17:51,733 POINTS OF CONTACTS IN IT WHERE 459 00:17:51,733 --> 00:17:54,436 THE TOPICS ALSO YOU CAN FIND 460 00:17:54,436 --> 00:17:57,406 POINTS OF MEDICAL COUNTER 461 00:17:57,406 --> 00:17:59,274 MEASURES.gov. 462 00:17:59,274 --> 00:18:02,311 WE OFTEN ENGAGE IN 463 00:18:02,311 --> 00:18:03,946 PRE-SOLICITATION CALLS TAILORED 464 00:18:03,946 --> 00:18:05,113 FOR THE AREA OF FROM OF THE 465 00:18:05,113 --> 00:18:06,548 BARDA PROGRAM. 466 00:18:06,548 --> 00:18:08,584 IF YOU HAVE A MORE GENERAL 467 00:18:08,584 --> 00:18:09,751 INQUIRY OR TECHNOLOGY, I 468 00:18:09,751 --> 00:18:13,956 ENCOURAGE YOU TO CHECK OUT OUR 469 00:18:13,956 --> 00:18:17,259 PROGRAM AND THIS IS AN ABILITY 470 00:18:17,259 --> 00:18:20,462 TO ENGAGE WITH ALL OF BARDA AS 471 00:18:20,462 --> 00:18:22,898 WELL AS OTHER U.S. GOVERNMENT 472 00:18:22,898 --> 00:18:23,232 PARTICIPANTS. 473 00:18:23,232 --> 00:18:25,601 YOU CAN GET DIRECT FEEDBACK ON 474 00:18:25,601 --> 00:18:29,071 YOUR TECHNOLOGY AND THAT'S AT 475 00:18:29,071 --> 00:18:32,007 MEDICAL COUNTER MEASURES.gov. 476 00:18:32,007 --> 00:18:33,742 SO, THIS IS WHERE YOU CAN 477 00:18:33,742 --> 00:18:34,943 CONNECT WITH US. 478 00:18:34,943 --> 00:18:36,712 I WANT TO PLUG BARDA INDUSTRY 479 00:18:36,712 --> 00:18:42,651 DAY WE ANNOUNCED THE DATES IT'S 480 00:18:42,651 --> 00:18:43,385 NEXT SUMMER. 481 00:18:43,385 --> 00:18:45,787 I ENCOURAGE YOU TO FOL US ON 482 00:18:45,787 --> 00:18:47,723 MEDICAL COUNTER MEASURES.gov FOR 483 00:18:47,723 --> 00:18:51,894 UPCOMING INFORMATION AND THANK 484 00:18:51,894 --> 00:18:53,161 YOU FOR BEING HERE AND 485 00:18:53,161 --> 00:18:54,229 PARTICIPATING IN THIS IMPORTANT 486 00:18:54,229 --> 00:18:58,300 MEETING AND HERE'S TO A 487 00:18:58,300 --> 00:18:59,668 PRODUCTIVE MEETING AND LOOK 488 00:18:59,668 --> 00:19:03,739 FORWARD TO LEARNING FROM YOU ALL 489 00:19:03,739 --> 00:19:04,339 TODAY. 490 00:19:04,339 --> 00:19:05,741 >> FANTASTIC, THANK YOU, KIM FOR 491 00:19:05,741 --> 00:19:07,276 THE WONDERFUL OPENING REMARKS 492 00:19:07,276 --> 00:19:09,745 AND LETTING US KNOW ALL ABOUT 493 00:19:09,745 --> 00:19:09,945 BARDA. 494 00:19:09,945 --> 00:19:13,081 OKAY, SO I'M GOING TO SHARE MY 495 00:19:13,081 --> 00:19:22,124 SLIDES NEXT. 496 00:19:22,124 --> 00:19:26,962 WHAT WE'LL DO IN THE NEXT 10 497 00:19:26,962 --> 00:19:28,564 MINUTES GO OVER THE WORKSHOP 498 00:19:28,564 --> 00:19:29,464 OBJECTIVES AND GOALS. 499 00:19:29,464 --> 00:19:33,735 I'D LIKE TO INTRODUCE MY 500 00:19:33,735 --> 00:19:39,741 COLLEAGUE, DR. ALEXEY ZAKHAROV 501 00:19:39,741 --> 00:19:41,944 AND HE'S CO-CLAIR AND 502 00:19:41,944 --> 00:19:44,146 INFORMATICS GROUP LEADER IN 503 00:19:44,146 --> 00:19:45,414 NCATS IN THE EARLY TRANSLATION 504 00:19:45,414 --> 00:19:48,917 BRANCH IN THE DIVISION OF PRE 505 00:19:48,917 --> 00:19:49,718 CLINICAL INNOVATION. 506 00:19:49,718 --> 00:19:56,825 SUBJECT MATTER EXPERT AND LEADS 507 00:19:56,825 --> 00:19:57,726 MANY PROJECTS. 508 00:20:00,896 --> 00:20:03,465 SO, SO WHAT I'LL DO IN THE NEXT 509 00:20:03,465 --> 00:20:06,034 TWO SLIDES, I'LL GO OVER THE 510 00:20:06,034 --> 00:20:09,004 PROGRAM AND ISLAND IT OVER TO 511 00:20:09,004 --> 00:20:10,205 ALEXEY AND HE'LL TALK TO YOU 512 00:20:10,205 --> 00:20:10,806 MORE ABOUT THE PARTICULAR 513 00:20:10,806 --> 00:20:11,073 WORKSHOP. 514 00:20:11,073 --> 00:20:13,742 OKAY. 515 00:20:13,742 --> 00:20:16,745 SO, AS JONI MENTIONED IN HER 516 00:20:16,745 --> 00:20:18,013 OPENING REMARKS, THE MISSION AND 517 00:20:18,013 --> 00:20:21,750 THE FOCUS OF MY PROGRAM IS TO 518 00:20:21,750 --> 00:20:26,188 TACKLE THE LACK OF 519 00:20:26,188 --> 00:20:26,955 REPRODUCIBILITY AND 520 00:20:26,955 --> 00:20:27,923 REPLICABILITY IN RESEARCH. 521 00:20:27,923 --> 00:20:31,293 AS MANY KNOW THERE'S A VERY HIGH 522 00:20:31,293 --> 00:20:35,297 RELEVANCE OF REPRODUCIBILITY IN 523 00:20:35,297 --> 00:20:36,431 PRECLINICAL RESEARCH AND DATA 524 00:20:36,431 --> 00:20:41,737 HAS SHOWN IN THE PAST DECADES IT 525 00:20:41,737 --> 00:20:42,537 EXCEEDS 50%. 526 00:20:42,537 --> 00:20:45,741 AND THIS IS ACTUALLY A STUDY WAS 527 00:20:45,741 --> 00:20:49,678 DONE THAT SHOWS THIS LEADS TO 528 00:20:49,678 --> 00:20:53,315 ABOUT $28 BILLION SPENT IN THE 529 00:20:53,315 --> 00:20:57,753 U.S. ALONE PER YEAR ON 530 00:20:57,753 --> 00:20:58,854 REPRODUCIBLE RESEARCH. 531 00:20:58,854 --> 00:21:01,456 AND IT INCLUDES BIOLOGICAL 532 00:21:01,456 --> 00:21:02,190 REAGENTS, MATERIALS, DATA DESIGN 533 00:21:02,190 --> 00:21:05,727 AND DATA ANALYSIS AND REPORTING 534 00:21:05,727 --> 00:21:06,428 AND PROTOCOLS. 535 00:21:06,428 --> 00:21:08,330 WHAT WE THINK IS HAPPENING IS 536 00:21:08,330 --> 00:21:10,365 THERE'S POOR TRAINING IN 537 00:21:10,365 --> 00:21:10,999 EXPERIMENTAL DESIGN IN DATA 538 00:21:10,999 --> 00:21:14,536 ANALYSIS AND PUBLICATIONS ARE 539 00:21:14,536 --> 00:21:16,071 NOT REPORTING BASIC ELEMENTS OF 540 00:21:16,071 --> 00:21:16,705 EXPERIMENTAL DESIGN. 541 00:21:16,705 --> 00:21:21,309 WHAT ARE THE RESOURCES WE OFFER? 542 00:21:21,309 --> 00:21:24,546 WE'RE HOPING TO ADVANCE THE 543 00:21:24,546 --> 00:21:27,182 PRACTICE AND VIGOR OF 544 00:21:27,182 --> 00:21:27,949 PRECLINICAL TRANSLATION AND HAVE 545 00:21:27,949 --> 00:21:29,351 THREE PLAIN COMPONENTS AND 546 00:21:29,351 --> 00:21:30,352 RESOURCES WE OFFER. 547 00:21:30,352 --> 00:21:33,722 THE FIRST IS THE MANUAL E-BOOK. 548 00:21:33,722 --> 00:21:35,390 A FREE BEST PRACTICES ONLINE 549 00:21:35,390 --> 00:21:38,727 RESOURCE DEVOTED TO THE 550 00:21:38,727 --> 00:21:41,663 SUCCESSFUL DEVELOPMENT OF EARLY 551 00:21:41,663 --> 00:21:43,565 DRUG DISCOVERY ASSAYS SUPPORTED 552 00:21:43,565 --> 00:21:47,202 BY 35 EDITORS ACROSS ALL SECTORS 553 00:21:47,202 --> 00:21:48,904 AVAILABLE FOR THOSE VIEWING 554 00:21:48,904 --> 00:21:51,807 ONLINE AND AT THE NATIONAL 555 00:21:51,807 --> 00:21:53,942 LIBRARY OF MEDICINE AND OPEN 556 00:21:53,942 --> 00:21:56,178 RESOURCE AND RUN WORKSHOPS AND 557 00:21:56,178 --> 00:21:57,746 WEBINARS SIMILAR TO IT TODAY AND 558 00:21:57,746 --> 00:22:03,085 THE OVER ALL GOAL IS TO 559 00:22:03,085 --> 00:22:04,786 INTRODUCE AND MOST OF THE 560 00:22:04,786 --> 00:22:06,555 WORKSHOP INCLUDING THIS ONE IS 561 00:22:06,555 --> 00:22:07,989 GOING TO BE RECORDED. 562 00:22:07,989 --> 00:22:11,727 IT'S BEING RECORDED AND 563 00:22:11,727 --> 00:22:13,729 AVAILABLE ONLINE FOR VIEWING 564 00:22:13,729 --> 00:22:14,463 AFTERWARDS. 565 00:22:14,463 --> 00:22:16,398 WE ALSO DO COLLABORATIVE 566 00:22:16,398 --> 00:22:18,300 RESEARCH FOCUSSED ON 567 00:22:18,300 --> 00:22:20,368 REPRODUCIBILITY AND ROBUSTNESS 568 00:22:20,368 --> 00:22:21,203 OF ASSAYS. 569 00:22:21,203 --> 00:22:24,439 I'M SHOWING A QR CODE TO THE 570 00:22:24,439 --> 00:22:24,906 RIGHT. 571 00:22:24,906 --> 00:22:26,274 IT WILL TAKE US TO THE WEBSITE 572 00:22:26,274 --> 00:22:28,310 AND IT'S A ONE-STOP SHOP YOU'LL 573 00:22:28,310 --> 00:22:32,814 GET ALL OF THESE RESOURCES THERE 574 00:22:32,814 --> 00:22:35,250 AFTER THE WORKSHOP. 575 00:22:35,250 --> 00:22:38,320 NOW I'LL PASS AND HAND IT OFF TO 576 00:22:38,320 --> 00:22:40,689 ALEXEY TO TALK MORE ABOUT THE 577 00:22:40,689 --> 00:22:41,156 PARTICULAR WORKSHOP. 578 00:22:41,156 --> 00:22:42,124 THE FLOOR YOURS. 579 00:22:42,124 --> 00:22:42,824 >> THANK YOU. 580 00:22:42,824 --> 00:22:44,192 I APPRECIATE IT. 581 00:22:44,192 --> 00:22:48,797 I'M EXCITED TO SEE A LOT OF 582 00:22:48,797 --> 00:22:56,204 PATIENTS OVER THE ZOOM AND NIH 583 00:22:56,204 --> 00:22:56,471 VIDEOCAST. 584 00:22:56,471 --> 00:23:07,015 OUR COMMITTEE HAS GOVERNMENT AND 585 00:23:12,988 --> 00:23:13,522 ACADEMIA. 586 00:23:13,522 --> 00:23:15,991 WE HAVE A SCIENTIFIC WHO MAKES 587 00:23:15,991 --> 00:23:18,360 THE WORKSHOP HAPPEN. 588 00:23:18,360 --> 00:23:19,961 AS WELL AS WE HAVE SUBJECT 589 00:23:19,961 --> 00:23:24,499 EXPERTS HERE AND CHIEF FROM 590 00:23:24,499 --> 00:23:29,771 BARDA AND AS WELL AS THE CHIEF 591 00:23:29,771 --> 00:23:40,148 FROM ACADEMIA AND UCSD. 592 00:23:40,148 --> 00:23:42,551 WE HAVE PARTICIPANTS. 593 00:23:42,551 --> 00:23:46,288 SO FAR WE HAVE AROUND 3,000 594 00:23:46,288 --> 00:23:47,355 REGISTRANTS FROM OVER 40 595 00:23:47,355 --> 00:23:47,622 COUNTRIES. 596 00:23:47,622 --> 00:23:51,226 WE HAVE AN ABOUT DISTRIBUTION. 597 00:23:51,226 --> 00:23:54,896 58 FROM ACADEMIA AND 10% FROM 598 00:23:54,896 --> 00:24:00,902 PHARMA AND 15% FROM BIO TECH AND 599 00:24:00,902 --> 00:24:03,505 12% FROM THE GOVERNMENT. 600 00:24:03,505 --> 00:24:05,740 NEXT TODAY AND TOMORROW WE'LL 601 00:24:05,740 --> 00:24:07,542 PURSUE THE FOLLOWING GOALS. 602 00:24:07,542 --> 00:24:10,645 WE WANT TO PROVIDE CONCEPTS AND 603 00:24:10,645 --> 00:24:12,681 APPROACH AND BEST PRACTICES TO 604 00:24:12,681 --> 00:24:15,417 SUCCESSFULLY DEVELOP AND 605 00:24:15,417 --> 00:24:19,487 INTEGRATE ROBUST AND RIGOROUS 606 00:24:19,487 --> 00:24:21,356 METHODS AND DISCUSSION IN TERMS 607 00:24:21,356 --> 00:24:22,791 OF UTILIZATION OF THE METHODS 608 00:24:22,791 --> 00:24:24,926 AND HOW THEY ADVANCE 609 00:24:24,926 --> 00:24:28,597 TRANSLATIONAL SCIENCE AND DRIVE 610 00:24:28,597 --> 00:24:29,764 DISCOVERY. 611 00:24:29,764 --> 00:24:33,468 WE WANT TO SHARE KNOWLEDGE. 612 00:24:33,468 --> 00:24:36,905 WE DON'T WANT TO MAKE THE 613 00:24:36,905 --> 00:24:38,773 INFORMATIC PEOPLE FOR INFORMATIC 614 00:24:38,773 --> 00:24:41,042 PEOPLE WE WANT TO TOUCH THE 615 00:24:41,042 --> 00:24:43,111 BROADER AUDIENCE INCLUDING 616 00:24:43,111 --> 00:24:47,282 CHEMISTRY AND BIOLOGISTS. 617 00:24:47,282 --> 00:24:49,985 WE WANT TO ADDRESS 618 00:24:49,985 --> 00:24:50,819 REPRODUCIBILITY AND 619 00:24:50,819 --> 00:24:51,920 CONSIDERATIONS FOR DEVELOPING 620 00:24:51,920 --> 00:24:54,923 NOT JUST ROBUST BUT REPRODUCIBLE 621 00:24:54,923 --> 00:25:05,233 IN SILICO MODELS. 622 00:25:07,602 --> 00:25:10,205 SECTION ONE IS TO BUILD MODELS. 623 00:25:10,205 --> 00:25:15,677 SESSION 2 IS THE METHODOLOGY AND 624 00:25:15,677 --> 00:25:15,977 ALGORITHMS. 625 00:25:15,977 --> 00:25:17,879 SESSION 3 APPLICATION OF IN 626 00:25:17,879 --> 00:25:23,051 SILICO METHODOLOGIES AND DRUG 627 00:25:23,051 --> 00:25:25,720 DISCOVERY PROFS AND SESSION 4 IS 628 00:25:25,720 --> 00:25:28,790 IN SILICO DRUG DISCOVERY AND THE 629 00:25:28,790 --> 00:25:30,458 PANEL DISCUSSION WHICH WILL 630 00:25:30,458 --> 00:25:31,893 BRIDGE THE GAPS IN 631 00:25:31,893 --> 00:25:32,227 TRANSLATIONAL. 632 00:25:32,227 --> 00:25:39,134 AS JONI MENTIONED TODAY, NCATS 633 00:25:39,134 --> 00:25:43,605 TRANSLATIONAL INSTITUTE LOOKS 634 00:25:43,605 --> 00:25:45,707 THAT THEY NOT ONLY TRANSLATE TO 635 00:25:45,707 --> 00:25:47,442 ANIMALS AND ANIMALS TO THE 636 00:25:47,442 --> 00:25:47,642 HUMAN. 637 00:25:47,642 --> 00:25:51,513 WE WANT TO FACILITATE DISCUSSION 638 00:25:51,513 --> 00:25:53,748 HOW A.I. AND MACHINE LEARNING 639 00:25:53,748 --> 00:26:00,221 CAN HELP BRIDGE THE GAPS. 640 00:26:00,221 --> 00:26:01,222 A BIT OF HOUSEKEEPING. 641 00:26:01,222 --> 00:26:03,925 HERE IS THE AGENDA FOR THE DAY. 642 00:26:03,925 --> 00:26:06,394 WE STARTED AND WE'RE GOING TO 643 00:26:06,394 --> 00:26:08,196 FINISH AT 5:00 P.M. 644 00:26:08,196 --> 00:26:11,866 WE'LL HAVE SEVERAL BREAKS AND 645 00:26:11,866 --> 00:26:14,536 LUNGE BREAKS AT 12:15. 646 00:26:14,536 --> 00:26:15,503 EACH SPEAKER WILL HAVE 15 647 00:26:15,503 --> 00:26:17,138 MINUTES FOR THE TALK AND FIVE 648 00:26:17,138 --> 00:26:18,907 MINUTES FOR QUESTIONS. 649 00:26:18,907 --> 00:26:23,111 WE HAVE AROUND 3,000 650 00:26:23,111 --> 00:26:24,913 PARTICIPANTS AND WE ENCOURAGE 651 00:26:24,913 --> 00:26:27,182 YOU TO TYPE A QUESTION IN THE 652 00:26:27,182 --> 00:26:28,416 Q&A PORTION OF THE ZOOM. 653 00:26:28,416 --> 00:26:31,853 IF YOU SEE A QUESTION YOU LIKE, 654 00:26:31,853 --> 00:26:34,622 PLEASE CLICK THE LIKE BUTTON SO 655 00:26:34,622 --> 00:26:38,360 THE MORE LIKES WILL GET THE 656 00:26:38,360 --> 00:26:39,361 PRIVATE TO BE ASKED. 657 00:26:39,361 --> 00:26:44,933 DAY 2 HAS A SIMILAR SCHEDULE 658 00:26:44,933 --> 00:26:46,368 FROM 11:00 A.M. TO 5:00 P.M. 659 00:26:46,368 --> 00:26:48,903 WITH SEVERAL BREAKS AND WE HAVE 660 00:26:48,903 --> 00:26:51,606 A UNIQUE SESSION. 661 00:26:51,606 --> 00:26:54,876 THE PANEL DISCUSSION AS WELL AS 662 00:26:54,876 --> 00:26:56,378 THE ADDITIONAL SESSION. 663 00:26:56,378 --> 00:26:58,646 AND WE ENCOURAGE EVERYONE WHO 664 00:26:58,646 --> 00:27:01,516 MAYBE ASKS A QUESTION OR COMES 665 00:27:01,516 --> 00:27:03,752 UP WITH A NEW QUESTION PLEASE 666 00:27:03,752 --> 00:27:05,720 TYPE THAT AND FACILITATE 667 00:27:05,720 --> 00:27:07,288 DISCUSSION IN THE TIME SLOT. 668 00:27:07,288 --> 00:27:10,325 WITH THAT SAID I WANT TO GIVE 669 00:27:10,325 --> 00:27:11,593 SARINE THE OPPORTUNITY TO 670 00:27:11,593 --> 00:27:12,227 INTRODUCE OUR KEY NOTE SPEAKER. 671 00:27:12,227 --> 00:27:22,170 THANK YOU. 672 00:27:22,170 --> 00:27:23,838 >> THANK YOU, ALEXEY. 673 00:27:23,838 --> 00:27:28,343 WE'RE SUPER EXCITED TO INTRODUCE 674 00:27:28,343 --> 00:27:30,912 OUR KEY NOTE SPEAKER, DR. MARTY 675 00:27:30,912 --> 00:27:31,579 HEAD. 676 00:27:31,579 --> 00:27:34,783 THE EXECUTIVE OF COMPUTATIONAL 677 00:27:34,783 --> 00:27:36,951 AND DATA SCIENCES AT AMGEN 678 00:27:36,951 --> 00:27:39,254 LEADING MECHANISTIC MACHINE 679 00:27:39,254 --> 00:27:40,755 LEARNING AND COMPUTATIONAL 680 00:27:40,755 --> 00:27:42,557 METHODS NO THE UNDERSTANDING OF 681 00:27:42,557 --> 00:27:44,926 DISEASE BIOLOGY AND THE 682 00:27:44,926 --> 00:27:52,200 DISCOVERY AND DESIGN OF BIOLOGIC 683 00:27:52,200 --> 00:27:55,603 AND SYNTHETIC THERAPEUTICS AND 684 00:27:55,603 --> 00:27:57,739 WORKED AT MULTIPLE INSTITUTIONS 685 00:27:57,739 --> 00:28:01,576 TODAY WHERE SHE WAS DIRECTOR OF 686 00:28:01,576 --> 00:28:02,610 COMPUTATIONAL BIO MEDICAL 687 00:28:02,610 --> 00:28:08,450 INITIATIVE AND OF THE BIOLOGICAL 688 00:28:08,450 --> 00:28:12,921 SCIENCES AND SPENT 20 YEARS IN 689 00:28:12,921 --> 00:28:16,257 BIOLOGY AT GLAXOSMITHKLINE AND 690 00:28:16,257 --> 00:28:22,297 WAS THE LEAD AT GSK AS THE TEAM 691 00:28:22,297 --> 00:28:26,401 AND WAS TO CREATIVELY APPLY ALL 692 00:28:26,401 --> 00:28:30,438 RELEVANT COMPUTATIONAL TOOLS TO 693 00:28:30,438 --> 00:28:31,439 PROGRESSING DRUG DISCOVERY 694 00:28:31,439 --> 00:28:33,408 EFFORTS FROM DRUG SELECTION TO 695 00:28:33,408 --> 00:28:36,911 THE CANDIDATE FOR CLINICAL 696 00:28:36,911 --> 00:28:37,145 TRIALS. 697 00:28:37,145 --> 00:28:37,946 MAR 698 00:28:37,946 --> 00:28:45,086 MARTY, WE LOOK FORWARD TO YOUR 699 00:28:45,086 --> 00:28:52,727 TALK. 700 00:28:52,727 --> 00:28:54,496 >> SO, IT'S A PLEASURE TO BE 701 00:28:54,496 --> 00:28:58,566 WITH ALL OF YOU TODAY. 702 00:28:58,566 --> 00:29:08,409 THANKS TO ROMI AND ALEXEY AND 703 00:29:08,409 --> 00:29:11,146 GLAD TO SEE I WAS READING YOUR 704 00:29:11,146 --> 00:29:15,283 MIND IN THINKING IING ABOUT THE 705 00:29:15,283 --> 00:29:18,887 TOPICS I WANTED TO SHARE AND 706 00:29:18,887 --> 00:29:19,487 THINKING ABOUT DRUG DISCOVERY 707 00:29:19,487 --> 00:29:26,895 THEN AND NOW. 708 00:29:26,895 --> 00:29:30,498 OF LATE WE SEE PAPER AFTER 709 00:29:30,498 --> 00:29:33,168 PAPER, PRESS RELEASE, PRESS 710 00:29:33,168 --> 00:29:34,135 RELEASE, PRESS RELEASE ABOUT 711 00:29:34,135 --> 00:29:36,905 A.I. AND DRUG DISCOVERY. 712 00:29:36,905 --> 00:29:45,446 IT FEELS TO LIKE TO ME WE HAVE 713 00:29:45,446 --> 00:29:47,415 HYPE TO GO AND IT'S A RANDOM 714 00:29:47,415 --> 00:29:49,717 SELECTION OF THINGS THAT WILL 715 00:29:49,717 --> 00:29:55,356 SHOW UP IN GOOGLE IF SCHOLAR IF 716 00:29:55,356 --> 00:29:57,725 YOU SEARCH FOR A.I. AND DRUG 717 00:29:57,725 --> 00:29:58,059 DISCOVERY. 718 00:29:58,059 --> 00:30:06,034 I COULD HAVE HAD HUNDREDS AND 719 00:30:06,034 --> 00:30:13,741 THOUSANDS MORE STAMS. 720 00:30:13,741 --> 00:30:16,511 I'D LIKE TO PAUSE AND TAKE A 721 00:30:16,511 --> 00:30:17,645 DEEP CALMING BREATH HOW WE COMB 722 00:30:17,645 --> 00:30:19,581 THROUGH IT TO IDENTIFY WAYS TO 723 00:30:19,581 --> 00:30:22,417 ACCELERATE THE HUNT FOR SAFE AND 724 00:30:22,417 --> 00:30:23,651 EFFECTIVE THERAPEUTICS THAT WILL 725 00:30:23,651 --> 00:30:24,919 MAKE A DIFFERENCE FOR PATIENTS 726 00:30:24,919 --> 00:30:31,125 AROUND THE WORLD. 727 00:30:31,125 --> 00:30:34,562 WE'RE GOING TO TALK ABOUT THE 728 00:30:34,562 --> 00:30:36,598 PREHISTORY AND MOVE ON TO 729 00:30:36,598 --> 00:30:38,132 THINKING ABOUT THE NOW AND THEN 730 00:30:38,132 --> 00:30:41,135 MAYBE A FEW THOUGHTS ABOUT HOW 731 00:30:41,135 --> 00:30:42,570 WE SHOULD APPROACH THE FUTURE 732 00:30:42,570 --> 00:30:50,378 FROM A PHILOSOPHICAL VIEW POINT. 733 00:30:50,378 --> 00:30:52,780 I HOPE SOME OF THE THINGS I 734 00:30:52,780 --> 00:30:53,848 SHARE TODAY WILL SET UP 735 00:30:53,848 --> 00:30:54,682 COLLEAGUES IN THE FIELD FOR 736 00:30:54,682 --> 00:30:55,283 THEIR TALKS IN THE COMING TWO 737 00:30:55,283 --> 00:31:00,054 DAYS. 738 00:31:00,054 --> 00:31:03,591 FIRST, I WANT TO POINT OUT THIS 739 00:31:03,591 --> 00:31:05,994 PAPER FROM 1937 ABOUT THE EFFECT 740 00:31:05,994 --> 00:31:10,231 OF STRUCTURE ON THE REACTIONS OF 741 00:31:10,231 --> 00:31:12,467 ORGANIC COMPOUNDS ESPECIALLY 742 00:31:12,467 --> 00:31:13,801 BENZENE DERIVATIVES. 743 00:31:13,801 --> 00:31:17,305 THIS IS THE WORK OF LOUIS HAMET 744 00:31:17,305 --> 00:31:23,578 LOOKING AT INITIALLY AS YOU 745 00:31:23,578 --> 00:31:34,088 CHANGE SUBSTITUENTS AND LEARN 746 00:31:38,293 --> 00:31:48,770 FROM THAT THE SAME SUBSTITUENTS 747 00:31:49,203 --> 00:31:53,474 GO TO OTHER PROPERTIES. 748 00:31:53,474 --> 00:31:55,176 I LIKE TO USE THIS EXAMPLE TO 749 00:31:55,176 --> 00:31:59,914 REMIND EVERYONE WE HAVE BEEN 750 00:31:59,914 --> 00:32:00,782 DOING MACHINE LEARNING SINCE 751 00:32:00,782 --> 00:32:01,749 BEFORE THERE WERE MACHINES TO 752 00:32:01,749 --> 00:32:03,751 LEARN ON. 753 00:32:03,751 --> 00:32:06,854 THERE'S NOTHING NEW UNDER THE 754 00:32:06,854 --> 00:32:08,690 SUN AND THERE'S MANY THINGS NEW 755 00:32:08,690 --> 00:32:09,490 UNDER THE SUN. 756 00:32:09,490 --> 00:32:12,460 WHAT IS IT THAT TOOK US FROM 757 00:32:12,460 --> 00:32:13,127 DOING MACHINE LEARNING ON GRAPH 758 00:32:13,127 --> 00:32:17,865 PAPER TO WHERE WE ARE NOW. 759 00:32:17,865 --> 00:32:22,670 YOU'LL ALSO NOTE THE METROPOLIS 760 00:32:22,670 --> 00:32:24,539 ALGORITHM MANY WORK WITH DAY IN 761 00:32:24,539 --> 00:32:28,509 AND DAY OUT THAT EXISTED IN 762 00:32:28,509 --> 00:32:32,280 THEORY AND IN THOUGHT AND IN 763 00:32:32,280 --> 00:32:33,715 PAINSTAKING PRACTICE LONG BEFORE 764 00:32:33,715 --> 00:32:39,620 THE COMING OF ANY ACT AND PLA 765 00:32:39,620 --> 00:32:43,925 PLAIN -- MANIAC I'M SHOWING HERE 766 00:32:43,925 --> 00:32:48,096 THE LEADING CUTTING-EDGE 767 00:32:48,096 --> 00:32:53,101 COMPUTER INSTALLED LOS ALAMOS 768 00:32:53,101 --> 00:32:54,902 AND THE COMPUTING ARCHITECTURE 769 00:32:54,902 --> 00:32:57,505 ENABLED THE ABILITY TO APPLY 770 00:32:57,505 --> 00:32:59,140 THIS ALGORITHM AND TO DO THIS 771 00:32:59,140 --> 00:33:03,911 WORK THAT LEADS TO SO MANY WORK 772 00:33:03,911 --> 00:33:12,286 IN METHODS EVEN NOW TODAY. 773 00:33:12,286 --> 00:33:14,155 THESE ARE CONCRETE EXAMPLES OF 774 00:33:14,155 --> 00:33:16,057 THE PAST AS COMPUTATIONAL 775 00:33:16,057 --> 00:33:16,624 CHEMISTS. 776 00:33:16,624 --> 00:33:18,393 I HAVEN'T EVEN NOTED AT THIS 777 00:33:18,393 --> 00:33:21,295 TIME, QUANTUM CHEMISTRY, 778 00:33:21,295 --> 00:33:25,733 MOLECULAR DYNAMICS, NEURAL NETS, 779 00:33:25,733 --> 00:33:28,136 ALL CONCEPTS DEVELOPED IN THOSE 780 00:33:28,136 --> 00:33:30,571 EARLY DAYS BEFORE THE ADVENT OF 781 00:33:30,571 --> 00:33:36,911 COMPUTERS BUT THEN ENABLED JUST 782 00:33:36,911 --> 00:33:41,716 REALLY A FLOURISHING OF NEW 783 00:33:41,716 --> 00:33:44,752 THOUGHTS AND IDEAS BECAUSE OF 784 00:33:44,752 --> 00:33:45,720 THE EXISTENCE OF THAT HARDWARE. 785 00:33:45,720 --> 00:33:47,789 HERE AT THE BEGINNING OF THIS 786 00:33:47,789 --> 00:33:49,690 TALK I'D LIKE TO OFFER EARLY 787 00:33:49,690 --> 00:33:49,991 REFLECTIONS. 788 00:33:49,991 --> 00:33:53,728 ONE OF THOSE IS THAT 789 00:33:53,728 --> 00:33:58,499 COMPUTATIONAL CHEMISTRY IS A 790 00:33:58,499 --> 00:34:00,601 DISCIPLINE WITH ITS OWN RICH 791 00:34:00,601 --> 00:34:03,337 HISTORY AND AREAS OF RESEARCH 792 00:34:03,337 --> 00:34:06,741 FOCUS THAT ARE WORTHY OF MINDS 793 00:34:06,741 --> 00:34:08,443 AND ATTENTION ESPECIALLY FOR 794 00:34:08,443 --> 00:34:12,914 THOSE WHO WORKED SO MANY YEARS 795 00:34:12,914 --> 00:34:14,515 IN SMALL MOLECULE DISCOVERY. 796 00:34:14,515 --> 00:34:17,718 WE FEEL LIKE THE NEW KIDS ON THE 797 00:34:17,718 --> 00:34:20,888 BLOCK RELATIVE TO ORGANIC, 798 00:34:20,888 --> 00:34:23,491 SYNTHETIC CHEMISTRY GOING ON FOR 799 00:34:23,491 --> 00:34:23,758 CENTURIES. 800 00:34:23,758 --> 00:34:28,563 BUT WE ARE A DISCIPLINE WITH A 801 00:34:28,563 --> 00:34:29,730 HISTORY, WITH A PRESENT AND 802 00:34:29,730 --> 00:34:30,264 FUTURE. 803 00:34:30,264 --> 00:34:33,668 I'D ALSO LIKE TO POINT OUT THESE 804 00:34:33,668 --> 00:34:36,270 NEW COMPUTATIONAL ARCHITECTURES 805 00:34:36,270 --> 00:34:37,638 OPENS UP NEW WAYS OF COMPUTING 806 00:34:37,638 --> 00:34:41,843 AND EVEN NEW WAYS OF THINKING. 807 00:34:41,843 --> 00:34:46,380 AND FINALLY YOU MAY NOT HAVE 808 00:34:46,380 --> 00:34:47,315 NOTICED IN THE PICTURE BEFORE 809 00:34:47,315 --> 00:34:52,487 BUT TWO OPERATORS OF MANIAC AND 810 00:34:52,487 --> 00:34:57,692 TWO AUTHORS ON THE METROPOLIS 811 00:34:57,692 --> 00:35:00,728 PAPER WERE WOMEN AND THEY HAVE 812 00:35:00,728 --> 00:35:04,899 BEEN IN COMPUTATIONAL CHEMISTRY 813 00:35:04,899 --> 00:35:05,733 SINCE THE EARLY DAYS AND FOR 814 00:35:05,733 --> 00:35:07,435 GROUPS AND COUNTRIES. 815 00:35:07,435 --> 00:35:09,737 WE'RE A DIVERSE COMMUNITY THAT 816 00:35:09,737 --> 00:35:15,109 COMES TOGETHER TO DO THIS WORK 817 00:35:15,109 --> 00:35:16,944 TO ATTACK THIS DISCIPLINE AND 818 00:35:16,944 --> 00:35:17,678 MAKE A DIFFERENCE. 819 00:35:17,678 --> 00:35:20,214 NOW WE'LL TRANSITION TO WHAT HAS 820 00:35:20,214 --> 00:35:21,716 CHANGED BETWEEN THEN AND NOW. 821 00:35:21,716 --> 00:35:24,018 I DON'T THINK I'M GOING TO STEAL 822 00:35:24,018 --> 00:35:27,321 THUNDER FROM STEVEN BURLEY BUT I 823 00:35:27,321 --> 00:35:29,657 AM GOING TO USE PDB AS MY 824 00:35:29,657 --> 00:35:33,728 EXAMPLE ON WHICH TO ANCHOR THIS 825 00:35:33,728 --> 00:35:44,205 TRANSITION FROM THEN TO NOW. 826 00:35:45,806 --> 00:35:47,608 WE CAN SEE THE PHENOMENAL GROWTH 827 00:35:47,608 --> 00:35:48,943 OF THE NUMBER OF CRYSTAL 828 00:35:48,943 --> 00:35:51,245 STRUCTURES AVAILABLE TO US TO 829 00:35:51,245 --> 00:35:55,917 LEARN FROM, TO TRAIN ON, TO USE 830 00:35:55,917 --> 00:35:58,519 FOR FUNDAMENTAL QUESTIONS ABOUT 831 00:35:58,519 --> 00:36:02,490 SPECIFIC PROTEIN SYSTEMS. 832 00:36:02,490 --> 00:36:05,526 SO, I'M GOING TO POINT OUT A FEW 833 00:36:05,526 --> 00:36:08,663 LANDMARKS ALONG THIS GRAPH. 834 00:36:08,663 --> 00:36:15,036 THE FIRST IS THE 1982 PAPER 835 00:36:15,036 --> 00:36:18,639 COMING OUT THEIR LABS OF TECH IN 836 00:36:18,639 --> 00:36:20,608 CLINICAL RESEARCH AT UCSF. 837 00:36:20,608 --> 00:36:22,443 THE THING THAT'S AMAZING TO ME 838 00:36:22,443 --> 00:36:28,149 ABOUT THIS IS THAT IN 1982, 839 00:36:28,149 --> 00:36:29,317 IRWIN AND SEVERAL OTHER PEOPLE 840 00:36:29,317 --> 00:36:31,519 FAMILIAR TO MANY OF US IN THE 841 00:36:31,519 --> 00:36:36,924 ROOM WERE ALREADY THINKING ABOUT 842 00:36:36,924 --> 00:36:41,195 PROTEIN LIGAND DOCKING AT THE 843 00:36:41,195 --> 00:36:43,331 TIME THERE WERE ONLY 117 PROTEIN 844 00:36:43,331 --> 00:36:47,902 STRUCTURES IN THE EDB. 845 00:36:47,902 --> 00:36:53,441 THE EARLY NASCENT EDB. 846 00:36:53,441 --> 00:36:56,877 IN ADDITION TO BEING ONE OF THE 847 00:36:56,877 --> 00:37:00,915 MOST GRACIOUS HUMAN BEINGS I'VE 848 00:37:00,915 --> 00:37:04,485 EVER HAD THE OPPORTUNITY TO MEET 849 00:37:04,485 --> 00:37:06,253 HE WAS PRESCIENT HOW THE 850 00:37:06,253 --> 00:37:07,154 COMPUTING TOOLS WITH THIS SMALL 851 00:37:07,154 --> 00:37:12,927 AMOUNT OF DATA AT THE TIME COULD 852 00:37:12,927 --> 00:37:14,629 LEAD TO BIG THINGS FOR THE 853 00:37:14,629 --> 00:37:17,732 FUTURE. 854 00:37:17,732 --> 00:37:22,503 I'LL GO ON TO MENTION THE CAS 855 00:37:22,503 --> 00:37:23,571 COMPETITIONS A LARGE EXPERIMENT 856 00:37:23,571 --> 00:37:26,507 TO PROACTIVELY AND PREDICTIVELY 857 00:37:26,507 --> 00:37:28,009 ASSESS THE PERFORMANCE OF 858 00:37:28,009 --> 00:37:29,777 ALGORITHMS FOR PREDICTING 859 00:37:29,777 --> 00:37:32,913 PROTEIN STRUCTURES. 860 00:37:32,913 --> 00:37:33,948 THIS IS THE EXACT TIME PERIOD 861 00:37:33,948 --> 00:37:38,419 WHEN I WAS FINISHING MY Ph.D. 862 00:37:38,419 --> 00:37:39,854 THINKING ABOUT GEOMETRIC 863 00:37:39,854 --> 00:37:41,555 PROPERTIES OF PROTEIN STRUCTURES 864 00:37:41,555 --> 00:37:49,730 AT DUKE UNIVERSITY WITH MICHAEL 865 00:37:49,730 --> 00:37:51,198 PRESONT AND MOVING TO THE CENTER 866 00:37:51,198 --> 00:37:52,833 FOR ADVANCED TECHNOLOGY FOR MY 867 00:37:52,833 --> 00:38:00,174 POSTDOC AND I SAT OUTSIDE THE 868 00:38:00,174 --> 00:38:03,277 OFFICES OF JOHN NOLT AND HAD THE 869 00:38:03,277 --> 00:38:05,279 OPPORTUNITY TO LEARN FROM HIM 870 00:38:05,279 --> 00:38:07,415 AND SEE HOW HE THOUGHT ABOUT THE 871 00:38:07,415 --> 00:38:09,717 PROACTIVE STUDIES OF THESE 872 00:38:09,717 --> 00:38:13,888 MODELS THAT WE'RE BUILDING 873 00:38:13,888 --> 00:38:15,289 ACROSS THE PARTITION. 874 00:38:15,289 --> 00:38:19,226 AND REALLY THINKING AHEAD OF 875 00:38:19,226 --> 00:38:21,529 TIME HOW CAN WE RIGOROUSLY 876 00:38:21,529 --> 00:38:22,897 UNDERSTAND HOW OUR ALGORITHMS 877 00:38:22,897 --> 00:38:27,201 ARE TRULY PERFORMING. 878 00:38:27,201 --> 00:38:30,905 MOVING FORWARD IN TIME, I CALL 879 00:38:30,905 --> 00:38:33,741 OUT IN THE EARLY 2000s WORK I'M 880 00:38:33,741 --> 00:38:36,544 HIGHLIGHTING HERE THE WORK I WAS 881 00:38:36,544 --> 00:38:38,079 MOST DIRECTLY ASSOCIATED WITH 882 00:38:38,079 --> 00:38:41,716 BUT AT THE SAME TIME, OUR 883 00:38:41,716 --> 00:38:44,418 FRIENDS AND COLLEAGUES LIKE KATE 884 00:38:44,418 --> 00:38:46,554 AND GEORGIA AND CHRISTOPHER AND 885 00:38:46,554 --> 00:38:47,588 OTHERS AT MERCK WERE DOING 886 00:38:47,588 --> 00:38:51,592 SIMILAR ANALYSES AND LOOKING 887 00:38:51,592 --> 00:38:54,028 CRITICALLY AT HOW THOSE DOCKING 888 00:38:54,028 --> 00:38:56,731 ALGORITHMS THAT HAD BEEN 889 00:38:56,731 --> 00:39:03,170 DEVELOPED OVER THE YEARS FROM 890 00:39:03,170 --> 00:39:05,740 1982 PUBLICATION, HOW WERE THEY 891 00:39:05,740 --> 00:39:07,141 ARE PERFORMING ON THE DRUG 892 00:39:07,141 --> 00:39:08,809 DISCOVERY PROGRAMS WE WERE DOING 893 00:39:08,809 --> 00:39:09,710 IN THE PHARMACEUTICAL INDUSTRY? 894 00:39:09,710 --> 00:39:14,682 HOW DO WE KNOW WHICH PROGRAM TO 895 00:39:14,682 --> 00:39:15,549 USE? 896 00:39:15,549 --> 00:39:17,318 HOW DO WE KNOW WHICH PREDICTIONS 897 00:39:17,318 --> 00:39:19,887 IT TAKE SERIOUSLY AND WHICH TO 898 00:39:19,887 --> 00:39:20,154 NOT? 899 00:39:20,154 --> 00:39:22,056 SO A CONTRIBUTION I THINK WE IN 900 00:39:22,056 --> 00:39:23,824 THE INDUSTRY MADE TO ACADEMIA 901 00:39:23,824 --> 00:39:26,660 AND TO SOFTWARE DEVELOPERS IN 902 00:39:26,660 --> 00:39:29,730 REALLY HAVING THAT KIND OF 903 00:39:29,730 --> 00:39:33,968 ROBUST TEST ON DATA THAT SAT 904 00:39:33,968 --> 00:39:36,370 BEHIND OUR PROPRIETARY BARRIERS. 905 00:39:36,370 --> 00:39:38,506 AND OF COURSE THAT LED THROUGH 906 00:39:38,506 --> 00:39:43,410 THE NIH AND THE FUNDING 907 00:39:43,410 --> 00:39:47,615 MECHANISMS THERE TO FIRST THE 908 00:39:47,615 --> 00:39:49,183 CESAR AND D3R RESOURCE THAT MADE 909 00:39:49,183 --> 00:39:53,721 SOME OF THAT DATA FROM INSIDE 910 00:39:53,721 --> 00:39:56,524 INDUSTRY CURATED AND AVAILABLE 911 00:39:56,524 --> 00:39:58,559 FOR MAKING BLIND PREDICTIONS FOR 912 00:39:58,559 --> 00:40:01,495 FOLKS WHO WERE NOT INSIDE OUR 913 00:40:01,495 --> 00:40:08,469 WALLS TO BE ABLE TO EXPAND THAT 914 00:40:08,469 --> 00:40:09,603 COMMUNITY OF PEOPLE TO USE THE 915 00:40:09,603 --> 00:40:11,572 RIGOROUS EVALUATIONS OF TOOLS WE 916 00:40:11,572 --> 00:40:12,940 WERE USING. 917 00:40:12,940 --> 00:40:14,708 AND OF COURSE WE ALREADY HEARD 918 00:40:14,708 --> 00:40:18,846 AND HIGHLIGHTED THE WORK OF 919 00:40:18,846 --> 00:40:21,715 ALPHA FOLD IN 2021 PERFORMING 920 00:40:21,715 --> 00:40:24,051 THE BEST IN THAT KAS COMPETITION 921 00:40:24,051 --> 00:40:27,254 THAT JOHN CREATED ALL THOSE 922 00:40:27,254 --> 00:40:28,322 YEARS AGO. 923 00:40:28,322 --> 00:40:35,729 CREATING THE ABILITY FOR US TO 924 00:40:35,729 --> 00:40:37,398 LEARN FROM THE STRUCTURES ALL 925 00:40:37,398 --> 00:40:40,634 THE TIME AND HAVE PREDICTIVE, 926 00:40:40,634 --> 00:40:41,569 PROSPECTIVE EVALUATION AT HAND 927 00:40:41,569 --> 00:40:46,874 TO TEST OUT AND SEE THE STEP 928 00:40:46,874 --> 00:40:48,642 CHANGE IN PERFORMANCE WHICH THEN 929 00:40:48,642 --> 00:40:53,681 LED OF COURSE TO THE NOBEL PRIZE 930 00:40:53,681 --> 00:40:54,682 EARLIER THIS YEAR. 931 00:40:54,682 --> 00:40:56,183 ONE THING I'D SAY TO YOU AT THE 932 00:40:56,183 --> 00:40:58,185 START OF YOUR CAREER NOW LOOK 933 00:40:58,185 --> 00:40:59,854 BEING AT THIS HISTORY, LOOKING 934 00:40:59,854 --> 00:41:03,090 AT THIS AMAZING WORK THAT'S 935 00:41:03,090 --> 00:41:05,726 BEING DONE BY THESE NOBEL 936 00:41:05,726 --> 00:41:07,595 LAUREATES AND OTHERS AROUND THE 937 00:41:07,595 --> 00:41:10,297 RESEARCH COMMUNITY, DON'T WORRY. 938 00:41:10,297 --> 00:41:12,199 THE WORK IS NOT DONE. 939 00:41:12,199 --> 00:41:15,636 THERE IS STILL MUCH TO DO EVEN 940 00:41:15,636 --> 00:41:18,939 IN UNDERSTANDING PROTEIN 941 00:41:18,939 --> 00:41:21,108 STRUCTURE. 942 00:41:21,108 --> 00:41:22,643 EVEN IN UNDERSTANDING BEYOND 943 00:41:22,643 --> 00:41:23,944 PROTEIN STRUCTURE TO PROTEIN 944 00:41:23,944 --> 00:41:25,679 DYNAMICS AND THE INTERACTIONS OF 945 00:41:25,679 --> 00:41:26,780 PROTEINS WITH OTHER THINGS. 946 00:41:26,780 --> 00:41:29,950 THERE IS STILL MUCH FOR US TO 947 00:41:29,950 --> 00:41:30,184 DO. 948 00:41:30,184 --> 00:41:34,955 SO I'M GOING TO TAKE A PAUSE 949 00:41:34,955 --> 00:41:37,291 HERE TO OFFER MORE REFLECTIONS 950 00:41:37,291 --> 00:41:40,728 BASED ON THAT EXAMPLE OF THE 951 00:41:40,728 --> 00:41:42,930 WORK THAT'S BEEN TO ON OVER THE 952 00:41:42,930 --> 00:41:43,130 YEARS. 953 00:41:43,130 --> 00:41:51,238 FIRST IS THAT THE PDB GREW TO 954 00:41:51,238 --> 00:41:56,477 ITS CURRENT SIZE BECAUSE OF A 955 00:41:56,477 --> 00:41:59,446 STEADY ACCRETION OF DATA OVER 956 00:41:59,446 --> 00:41:59,647 TIME. 957 00:41:59,647 --> 00:42:01,649 THERE'S A ROUGHLY GEOMETRIC 958 00:42:01,649 --> 00:42:03,817 PROGRESSION IN THE GROWTH OF THE 959 00:42:03,817 --> 00:42:06,320 NUMBERS OF STRUCTURES DEPOSITED 960 00:42:06,320 --> 00:42:06,954 EACH YEAR. 961 00:42:06,954 --> 00:42:12,726 BECAUSE THE REPOSITORY EXISTS 962 00:42:12,726 --> 00:42:14,762 FOR US THAT DATA HAS GROWN OVER 963 00:42:14,762 --> 00:42:16,330 TIME TO APPROACH 250,000 964 00:42:16,330 --> 00:42:16,931 STRUCTURES THAT ARE AVAILABLE 965 00:42:16,931 --> 00:42:25,239 NOW. 966 00:42:25,239 --> 00:42:27,574 MOST WERE GENERATED FOR PRIMARY 967 00:42:27,574 --> 00:42:28,742 PURPOSE BECAUSE SOMEONE HAD A 968 00:42:28,742 --> 00:42:30,878 SPECIFIC INTEREST IN A 969 00:42:30,878 --> 00:42:33,714 PARTICULAR PROTEIN. 970 00:42:33,714 --> 00:42:34,915 FOR A SPECIFIC DISEASE 971 00:42:34,915 --> 00:42:36,216 INDICATION, FOR A SPECIFIC 972 00:42:36,216 --> 00:42:40,688 FAMILY OF PROTEINS THAT THEY 973 00:42:40,688 --> 00:42:41,322 WANTED TO UNDERSTAND HERE'S 974 00:42:41,322 --> 00:42:43,624 WHERE I CALL OUT STRUCTURES FROM 975 00:42:43,624 --> 00:42:47,361 THE EARLY DAYS THAT SOME PLAY 976 00:42:47,361 --> 00:42:50,864 REMEMBER WHERE THERE WERE 977 00:42:50,864 --> 00:42:52,399 LITERALLY HUNDREDS OF STRUCTURES 978 00:42:52,399 --> 00:42:55,569 WITH SMALL CHANGES BEING MADE TO 979 00:42:55,569 --> 00:42:56,804 CREATE POCKETS IN THE INTERIOR 980 00:42:56,804 --> 00:42:57,705 MUCH THE PROTEIN. 981 00:42:57,705 --> 00:43:01,008 RAT ONE TIME, A BIG PROPORTION 982 00:43:01,008 --> 00:43:03,677 OF THE STRUCTURES IN THE PDB 983 00:43:03,677 --> 00:43:06,981 WERE OF THIS ONE PROTEIN WITH 984 00:43:06,981 --> 00:43:12,686 THESE STRUCTURES CREATED TO ASK 985 00:43:12,686 --> 00:43:15,055 QUESTIONS OF PHYSICAL CHEMISTRY 986 00:43:15,055 --> 00:43:18,425 ABOUT THE STRUSHGZ AND CALLING 987 00:43:18,425 --> 00:43:22,696 THAT OUT BECAUSE BOTH THE 988 00:43:22,696 --> 00:43:24,999 PRIME -- STRUCTURES AND CALLING 989 00:43:24,999 --> 00:43:30,971 THAT OUT BECAUSE THE USE OF THE 990 00:43:30,971 --> 00:43:36,477 STRUCTURES ARE BOTH IMPORTANT. 991 00:43:36,477 --> 00:43:40,881 AND THE IMPORTANCE OF DATA 992 00:43:40,881 --> 00:43:43,017 QUANTITY, DATA QUALITY, DATA 993 00:43:43,017 --> 00:43:49,757 INTEGR 994 00:43:49,757 --> 00:43:50,324 INTEGRITY WHERE I CALL OUT 995 00:43:50,324 --> 00:43:53,494 PIONEERS WHO SAT WITH ME ALL 996 00:43:53,494 --> 00:43:56,663 THOSE YEARS AGO AND ALL THE REST 997 00:43:56,663 --> 00:43:58,665 OF THE PDB CREW WHO WERE 998 00:43:58,665 --> 00:44:03,837 PRESCIENT IN CREATING A ROBUST 999 00:44:03,837 --> 00:44:07,007 METAONTOLOGY TO DESCRIBE PROTEIN 1000 00:44:07,007 --> 00:44:08,375 STRUCTURES EVEN BACK TO THE 1001 00:44:08,375 --> 00:44:11,645 ORIGINAL PDB FORMAT THAT WAS 1002 00:44:11,645 --> 00:44:13,747 TRAPPED IN A CERTAIN NUMBER OF 1003 00:44:13,747 --> 00:44:19,620 CHARACTERS ON EACH LINE TO THE 1004 00:44:19,620 --> 00:44:22,689 FORMATS THAT ENABLED MORE 1005 00:44:22,689 --> 00:44:24,725 FLEXIBILITY IN THE METADATA. 1006 00:44:24,725 --> 00:44:27,628 THE TENACITY IN ENSURING THAT 1007 00:44:27,628 --> 00:44:29,563 THEY IMPOSE STANDARDS SO THAT 1008 00:44:29,563 --> 00:44:32,232 WHEN YOU SUBMIT A STRUCTURE TO 1009 00:44:32,232 --> 00:44:36,770 THE PDB, IT HAS THE DATA QUALITY 1010 00:44:36,770 --> 00:44:41,742 BAKED INTO IT AND HANGING OUT 1011 00:44:41,742 --> 00:44:49,750 THAT CARROT OF THE PUBLICATION 1012 00:44:49,750 --> 00:44:52,186 PUBLISHERS HAVING THE POLICY 1013 00:44:52,186 --> 00:44:54,421 THAT YOU MUST SUBMIT AND GET A 1014 00:44:54,421 --> 00:44:55,889 PDB CODE BEFORE YOU CAN PUBLISH 1015 00:44:55,889 --> 00:44:57,124 THE STRUCTURE. 1016 00:44:57,124 --> 00:44:58,892 THAT COLLECTION OF THINGS 1017 00:44:58,892 --> 00:45:01,595 ENSURED THAT THE DATA THAT IS IN 1018 00:45:01,595 --> 00:45:04,598 THERE IS RELIABLE FOR US TO USE 1019 00:45:04,598 --> 00:45:07,000 TO LEARN FROM AND EVEN MORE 1020 00:45:07,000 --> 00:45:08,969 IMPORTANTLY THE DECISIONS SORT 1021 00:45:08,969 --> 00:45:12,573 OF IN THE MID LIFE OF THE PDB TO 1022 00:45:12,573 --> 00:45:14,007 ENSURE STRUCTURE FACTORS WERE 1023 00:45:14,007 --> 00:45:16,043 INCLUDED AS WELL SO WE CAN GO 1024 00:45:16,043 --> 00:45:19,746 BACK AND REFINE STRUCTURES THAT 1025 00:45:19,746 --> 00:45:21,748 HAVE QUESTIONABLE THINGS AND ASK 1026 00:45:21,748 --> 00:45:23,517 METHODS OF IMPROVEMENTS. 1027 00:45:23,517 --> 00:45:25,052 THE FINAL REFLECTION I WANT TO 1028 00:45:25,052 --> 00:45:26,653 OFFER ABOUT THIS EXAMPLE I 1029 00:45:26,653 --> 00:45:30,991 SHARED WITH YOU IS THAT EACH OF 1030 00:45:30,991 --> 00:45:36,964 THOSE 250,000 STRUCTURES IS A 1031 00:45:36,964 --> 00:45:38,332 DYNAMICISH SNAPSHOT OF A 1032 00:45:38,332 --> 00:45:44,238 PARTICULAR STATE OF A PARTICULAR 1033 00:45:44,238 --> 00:45:49,009 PROTEIN CONSTRUCT. 1034 00:45:49,009 --> 00:45:54,448 AND SITTING IN A STRUCTURE FORM 1035 00:45:54,448 --> 00:45:56,383 AND THE CHALLENGE OF HAVING 1036 00:45:56,383 --> 00:45:57,718 MULTIPLE MODELS THAT MAP TO THE 1037 00:45:57,718 --> 00:46:00,521 DATA YOU SEE IN THE NMR 1038 00:46:00,521 --> 00:46:04,558 EXPERIMENTS AND IN CRYO-EM THE 1039 00:46:04,558 --> 00:46:07,027 TEMPERATURES AND MATRICES AND 1040 00:46:07,027 --> 00:46:10,264 GENERATING A STRUCTURE. 1041 00:46:10,264 --> 00:46:13,967 IT IS ALL A SNAPSHOT THAT HAS 1042 00:46:13,967 --> 00:46:17,738 ENCAPSULATES SOME OF THE MOTION 1043 00:46:17,738 --> 00:46:19,039 WE KNOW IS THERE BUT ONLY ONE 1044 00:46:19,039 --> 00:46:21,742 PICTURE OF WHAT THAT STRUCTURE 1045 00:46:21,742 --> 00:46:22,643 TRULY IS. 1046 00:46:22,643 --> 00:46:25,746 THIS IS MY OPPORTUNITY TO REMIND 1047 00:46:25,746 --> 00:46:29,383 ALL OF US AGAIN THAT EVEN WET 1048 00:46:29,383 --> 00:46:31,952 LAB EXPERIMENTS ARE A MODEL OF 1049 00:46:31,952 --> 00:46:37,591 THE THING THAT WE REALLY WANT TO 1050 00:46:37,591 --> 00:46:41,295 UNDERS 1051 00:46:41,295 --> 00:46:41,762 UNDERSTAND. 1052 00:46:41,762 --> 00:46:44,097 WE OFTEN CUT OFF THE KINASE 1053 00:46:44,097 --> 00:46:45,399 DOMAIN WITHOUT THINKING ABOUT 1054 00:46:45,399 --> 00:46:46,266 THE OTHER DOMAINS. 1055 00:46:46,266 --> 00:46:48,435 WE THINK ABOUT THEM, WE DO, 1056 00:46:48,435 --> 00:46:50,370 WE'RE CAREFUL SCIENTISTS BUT NOT 1057 00:46:50,370 --> 00:46:51,405 HAVING THE DATA THAT NECESSARY 1058 00:46:51,405 --> 00:46:52,773 TELLS US WHAT'S GOING ON WITH 1059 00:46:52,773 --> 00:46:54,308 THE EXTRA DOMAINS. 1060 00:46:54,308 --> 00:46:57,277 SO YES, A REMINDER FOR US ALL 1061 00:46:57,277 --> 00:47:00,447 THAT EVEN WET LAB EXPERIMENTS 1062 00:47:00,447 --> 00:47:02,716 ARE MODEL SYSTEMS OF THE BIOLOGY 1063 00:47:02,716 --> 00:47:04,451 THAT WE CARE ABOUT THAT GOES ON 1064 00:47:04,451 --> 00:47:08,255 WITH PATIENTS WHO HAVE DISEASE. 1065 00:47:08,255 --> 00:47:10,123 AND SO NOW WE'RE GOING TO MOVE 1066 00:47:10,123 --> 00:47:11,692 ON TO -- SO WHAT HAS CHANGED 1067 00:47:11,692 --> 00:47:12,492 SINCE THEN? 1068 00:47:12,492 --> 00:47:16,129 WE TALKED ABOUT THE GROWTH OF 1069 00:47:16,129 --> 00:47:16,330 DATA. 1070 00:47:16,330 --> 00:47:19,032 WE TALKED ABOUT HOW COMPUTING 1071 00:47:19,032 --> 00:47:22,102 WAS ORIGINALLY A TRANSFORMATIVE 1072 00:47:22,102 --> 00:47:25,038 ACT THAT ENABLED NEW WAYS OF 1073 00:47:25,038 --> 00:47:27,507 COMPUTING AND THINKING. 1074 00:47:27,507 --> 00:47:29,810 MY FATHER, MANY YEARS AGO BUILT 1075 00:47:29,810 --> 00:47:35,649 OUR FIRST COMPUTER. 1076 00:47:35,649 --> 00:47:38,418 OLD 8088 SITTING THAT LOOKED 1077 00:47:38,418 --> 00:47:39,953 LIKE THIS HERE AND MY PHONE 1078 00:47:39,953 --> 00:47:41,221 WHICH I'M KEEPING TRACK OF THE 1079 00:47:41,221 --> 00:47:44,591 TIME I'VE BEEN SPEAKING, HAS 1080 00:47:44,591 --> 00:47:46,493 MORE POWER THAT THAT COMPUTER 1081 00:47:46,493 --> 00:47:49,029 THAT SAT ON THE DESK IN OUR 1082 00:47:49,029 --> 00:47:58,505 HOUSE IN FLORIDA. 1083 00:47:58,505 --> 00:48:00,207 AND COMPETING ARCHITECTURE HAS 1084 00:48:00,207 --> 00:48:05,779 GROWN TO SYSTEMS LIKE THE N 1085 00:48:05,779 --> 00:48:09,750 VIDDIA DGX BOXES AND INSTALLING 1086 00:48:09,750 --> 00:48:13,053 FRONTIER, THE WORLD'S FASTER 1087 00:48:13,053 --> 00:48:23,330 SUPER COMPUTER. 1088 00:48:24,865 --> 00:48:27,267 THERE'S CAPABILITIES AVAILABLE 1089 00:48:27,267 --> 00:48:33,707 TO US AT THE TOUCH OF A COMPUTER 1090 00:48:33,707 --> 00:48:36,276 AND INCLUDING THE CLOUD 1091 00:48:36,276 --> 00:48:37,044 COMPUTING TIME WE HAVE A LOT 1092 00:48:37,044 --> 00:48:39,012 AVAILABLE TO US. 1093 00:48:39,012 --> 00:48:40,647 AS WE THINK TO THE FUTURE THERE 1094 00:48:40,647 --> 00:48:43,417 ARE NEW ARCHITECTURES COMING. 1095 00:48:43,417 --> 00:48:47,087 JONI MENTIONED QUANTUM 1096 00:48:47,087 --> 00:48:48,522 COMPUTING. 1097 00:48:48,522 --> 00:48:53,427 THERE'S A COURSE NEURO MORPHIC 1098 00:48:53,427 --> 00:48:56,229 COMPUTING NEURO PHYSICS AND IT 1099 00:48:56,229 --> 00:48:57,764 TAKES OUR UNDERSTANDING OF THE 1100 00:48:57,764 --> 00:49:00,734 BRAIN AS A WAY OF THINKING ABOUT 1101 00:49:00,734 --> 00:49:01,768 COMPUTE ARCHITECTURES FOR THE 1102 00:49:01,768 --> 00:49:10,777 FUTURE. 1103 00:49:10,777 --> 00:49:13,180 AND THERE'S UNTOLD THOUGHTS FOR 1104 00:49:13,180 --> 00:49:18,752 THE FUTURE NO ONE HAS DREAMT OF 1105 00:49:18,752 --> 00:49:20,020 YET OPENING UP NEW OPPORTUNITIES 1106 00:49:20,020 --> 00:49:21,021 FOR COMPUTING AND I'D LIKE TO 1107 00:49:21,021 --> 00:49:21,755 CHALLENGE YOU TO THINK ABOUT 1108 00:49:21,755 --> 00:49:24,424 WHAT IS THE NEW WAY OF THINKING 1109 00:49:24,424 --> 00:49:29,763 THAT THESE NEW ARCHITECTURES CAN 1110 00:49:29,763 --> 00:49:32,032 UNLEASH? 1111 00:49:32,032 --> 00:49:33,700 NOT JUST HOW CAN I TAKE 1112 00:49:33,700 --> 00:49:36,269 CLASSICAL ALGORITHMS AND 1113 00:49:36,269 --> 00:49:37,571 HOPEFULLY MAKE THEM RUN FASTER 1114 00:49:37,571 --> 00:49:41,041 ON THE VERY EXPENSIVE, VERY 1115 00:49:41,041 --> 00:49:43,043 ENERGY HUNGRY QUANTUM COMPUTERS 1116 00:49:43,043 --> 00:49:45,679 WE'RE BUILDING, BUT HOW DOES 1117 00:49:45,679 --> 00:49:50,584 THAT NEW WAY OF MOVING FROM BITS 1118 00:49:50,584 --> 00:49:53,687 TO KU 1119 00:49:56,723 --> 00:49:59,493 KUBITZ DEVELOP NEW INSIGHT AND 1120 00:49:59,493 --> 00:50:01,027 ACHIEVE TRANSLATIONAL VISIONS 1121 00:50:01,027 --> 00:50:11,304 ALL OF US HAVE? 1122 00:50:15,242 --> 00:50:19,045 I'M GOING TO TELL YOU A STORY 1123 00:50:19,045 --> 00:50:21,748 FROM THE SECOND HALF OF THE LAST 1124 00:50:21,748 --> 00:50:22,883 DECADE FROM THE EARLY DAYS WHEN 1125 00:50:22,883 --> 00:50:25,018 THE PUBLICATIONS STARTED SHOWING 1126 00:50:25,018 --> 00:50:28,088 UP IN THE RAPID FIRE OF FRENZY 1127 00:50:28,088 --> 00:50:28,955 OF PUBLICATIONS I SHARED WITH 1128 00:50:28,955 --> 00:50:30,257 YOU EARLY. 1129 00:50:30,257 --> 00:50:32,259 I WANT TO THINK ABOUT USING THE 1130 00:50:32,259 --> 00:50:39,266 EXAMPLE NOT TO SAY IT IS THE BE 1131 00:50:39,266 --> 00:50:42,769 ALL AND END ALL BUT WHAT CAN WE 1132 00:50:42,769 --> 00:50:45,505 LEARN FROM THE EARLY ERA OF A.I. 1133 00:50:45,505 --> 00:50:46,773 AND DRUG DISCOVERY THAT HELPED 1134 00:50:46,773 --> 00:50:53,346 TEACH US WHAT WE NEED TO DO NEXT 1135 00:50:53,346 --> 00:50:57,117 TO MAKE SURE THESE METHODS ARE 1136 00:50:57,117 --> 00:50:59,686 AS IMPACTFUL AS POSSIBLE. 1137 00:50:59,686 --> 00:51:03,690 I'LL TALK TO YOU ABOUT A 1138 00:51:03,690 --> 00:51:04,991 PUBLIC-PRIVATE PARTNERSHIP AT MY 1139 00:51:04,991 --> 00:51:15,502 TIME RAT GSK AND THE 1140 00:51:15,969 --> 00:51:17,737 PUBLIC-PRIVATE PARTNERSHIP UNDER 1141 00:51:17,737 --> 00:51:20,440 THE AUSPICES OF THEN VICE 1142 00:51:20,440 --> 00:51:25,011 PRESIDENT BIDEN'S CANCER 1143 00:51:25,011 --> 00:51:26,813 MOONSHOT. 1144 00:51:26,813 --> 00:51:31,985 THIS WAS A PROJECT AIMED AT 1145 00:51:31,985 --> 00:51:35,522 CREATING THE CLOSED ACTIVE 1146 00:51:35,522 --> 00:51:38,458 LEARNING GROUP WHERE WE HAD 1147 00:51:38,458 --> 00:51:41,495 PROPERTY FEEDING INTO GENERATIVE 1148 00:51:41,495 --> 00:51:43,563 MOLECULAR DESIGN ALGORITHMS AND 1149 00:51:43,563 --> 00:51:49,002 FEEDING THAT DATA INTO ACTIVE 1150 00:51:49,002 --> 00:51:54,574 LEARNING MECHANISMS WHEN WE SEE 1151 00:51:54,574 --> 00:51:57,744 OUR ACTIVE MODELS DRIFTING OUT 1152 00:51:57,744 --> 00:51:59,379 OF CONTROLS CAN WE FILL DESIGNS 1153 00:51:59,379 --> 00:52:01,348 TO FILL THE GAPS IN THE SPACES 1154 00:52:01,348 --> 00:52:03,283 WHERE COMPUTING WAS AGAINST. 1155 00:52:03,283 --> 00:52:04,951 CAN WE TRIGGER DESIGNS THAT MOVE 1156 00:52:04,951 --> 00:52:08,722 US TOWARDS THE MOLECULES THAT WE 1157 00:52:08,722 --> 00:52:09,689 NEED TO BE DRUGS. 1158 00:52:09,689 --> 00:52:11,558 SO MANY OF YOU WILL HAVE SEEN 1159 00:52:11,558 --> 00:52:13,827 MANY VERSIONS OF THIS KIND OF 1160 00:52:13,827 --> 00:52:19,633 THING IN THE YEARS SINCE WE 1161 00:52:19,633 --> 00:52:20,800 BEGAN AGGRESSIVELY APPLYING 1162 00:52:20,800 --> 00:52:22,836 THESE KINDS OF TECHNIQUES TO 1163 00:52:22,836 --> 00:52:25,038 DRUG DISCOVERY AND 1164 00:52:25,038 --> 00:52:26,172 COMPUTER-AIDED DRUG DESIGN. 1165 00:52:26,172 --> 00:52:29,509 WE DID TWO PILOT STUDIES AS PART 1166 00:52:29,509 --> 00:52:31,044 OF THIS PROJECT. 1167 00:52:31,044 --> 00:52:34,748 ONE HISTORICAL EXAMPLE FROM 1168 00:52:34,748 --> 00:52:38,518 KINASE INHIBITORS AT GSK. 1169 00:52:38,518 --> 00:52:44,324 THE SECOND SELECTIVE H1 1170 00:52:44,324 --> 00:52:46,526 ANTAGONIST WITH A COMPANY IN SAN 1171 00:52:46,526 --> 00:52:46,726 DIEGO. 1172 00:52:46,726 --> 00:52:48,495 SIMILAR DESIGN CRITERIA SHARED 1173 00:52:48,495 --> 00:52:52,432 BETWEEN THE TWO WITH SOME 1174 00:52:52,432 --> 00:52:55,902 SPECIFIC DESIGN CRITERIA FOR 1175 00:52:55,902 --> 00:52:58,371 EACH INDIVIDUAL ONES AND 1176 00:52:58,371 --> 00:53:03,944 SPECIFIC QUALITIESES -- 1177 00:53:03,944 --> 00:53:05,078 QUALITIES AND QUANTITIES WE HAD 1178 00:53:05,078 --> 00:53:07,681 TO WORK WITH AND DID A NUMBER OF 1179 00:53:07,681 --> 00:53:09,649 STUDIES DOING GENERATIVE DESIGN, 1180 00:53:09,649 --> 00:53:14,087 MAKING MOLECULES, TESTING IT AND 1181 00:53:14,087 --> 00:53:18,959 DOING A PROSPECTIVE, PREDICTIVE 1182 00:53:18,959 --> 00:53:21,061 ASSESSMENT AND USING IT AND 1183 00:53:21,061 --> 00:53:26,666 LEARNING FROM IT IN ACTION 1184 00:53:26,666 --> 00:53:28,802 BUILDING THE PLANE WHILE WE'RE 1185 00:53:28,802 --> 00:53:30,704 FLYING AND LEARNING WHAT WORKS 1186 00:53:30,704 --> 00:53:31,638 WHILE IT'S FLYING. 1187 00:53:31,638 --> 00:53:34,474 WHAT CAN WE LEARN FROM THE ONE 1188 00:53:34,474 --> 00:53:36,042 EARLY EXPERIMENT IN THESE KINDS 1189 00:53:36,042 --> 00:53:36,576 OF METHODS? 1190 00:53:36,576 --> 00:53:38,278 THE FIRST THING WE CAN AND DID 1191 00:53:38,278 --> 00:53:44,317 LEARN IS GENERATIVE DESIGN AND 1192 00:53:44,317 --> 00:53:45,085 MULTI-PARAMETER OPTIMIZATION 1193 00:53:45,085 --> 00:53:47,320 SHIFTS THE DISTRIBUTION OF 1194 00:53:47,320 --> 00:53:50,790 PREDICTIVE AND MEASURED 1195 00:53:50,790 --> 00:53:53,093 PROPERTIES IN THE DESIRED 1196 00:53:53,093 --> 00:53:53,360 DIRECTION. 1197 00:53:53,360 --> 00:53:55,061 SO, JUST A COUPLE OF EXAMPLE OF 1198 00:53:55,061 --> 00:53:56,429 GRAPHS OF PAPERS FROM COLLEAGUES 1199 00:53:56,429 --> 00:54:01,034 AT THAT TIME LOOKING AT THE 1200 00:54:01,034 --> 00:54:04,070 AURORA KINASE EXAMPLE WHERE THE 1201 00:54:04,070 --> 00:54:06,606 TRAINING DATA AND THE FIRST SIX 1202 00:54:06,606 --> 00:54:09,576 MONTHS DATA THAT IS MEASURED IS 1203 00:54:09,576 --> 00:54:11,111 GRAPHED HERE AND THE BEST 250 1204 00:54:11,111 --> 00:54:14,180 OVER THOSE ITERATIONS, YOU CAN 1205 00:54:14,180 --> 00:54:18,585 SEE WE HAVE MOVED THE PROPERTIES 1206 00:54:18,585 --> 00:54:20,186 OF SELECTIVITY IN THE DIRECTION 1207 00:54:20,186 --> 00:54:23,923 WE WANTED AND ALSO BEGAN 1208 00:54:23,923 --> 00:54:24,824 SHIFTING SOME OF THE 1209 00:54:24,824 --> 00:54:25,759 PROPERTIES -- THE OTHER 1210 00:54:25,759 --> 00:54:27,360 PROPERTIES THAT GO INTO MAKING A 1211 00:54:27,360 --> 00:54:29,429 DRUG THAT WE CARE ABOUT AS WELL. 1212 00:54:29,429 --> 00:54:31,464 OTHER SNAPSHOTS TO LET YOU SEE 1213 00:54:31,464 --> 00:54:34,601 THIS IN ACTION AS WE GO THROUGH 1214 00:54:34,601 --> 00:54:37,737 ITERATIONS OF THE GENERATIVE 1215 00:54:37,737 --> 00:54:40,106 DESIGN ALGORITHM SEEING THE 1216 00:54:40,106 --> 00:54:41,741 DESIGN MOLECULES PULL AWAY FROM 1217 00:54:41,741 --> 00:54:45,745 THE SPACE OF THE ONES WE STARTED 1218 00:54:45,745 --> 00:54:50,850 WITH BOTH THERE AND IN THIS 1219 00:54:50,850 --> 00:54:51,084 PICTURE. 1220 00:54:51,084 --> 00:54:53,953 AND SO THEY ABSOLUTELY TAKE THAT 1221 00:54:53,953 --> 00:54:57,023 WHOLE DISTRIBUTION AND MOVE IT 1222 00:54:57,023 --> 00:54:58,058 IN THE RIGHT DIRECTION. 1223 00:54:58,058 --> 00:55:01,327 HOWEVER, IT IS ABSOLUTELY TRUE 1224 00:55:01,327 --> 00:55:03,663 THAT SPECIFIC PREDICTIONS FOR 1225 00:55:03,663 --> 00:55:05,732 SPECIFIC MOLECULES ARE VERY 1226 00:55:05,732 --> 00:55:07,534 UNLIKELY TO BE ACCURATE. 1227 00:55:07,534 --> 00:55:11,004 AND SO IF WE TREAT THE NUMBER 1228 00:55:11,004 --> 00:55:13,039 FOR SOLUBILITY AS A PREDICTION 1229 00:55:13,039 --> 00:55:15,275 OF WHAT IT WILL BE, WE'RE USING 1230 00:55:15,275 --> 00:55:16,676 IT IN THE WRONG WAY. 1231 00:55:16,676 --> 00:55:18,311 IT IS NOT A PREDICTION OF THE 1232 00:55:18,311 --> 00:55:18,978 NUMBER. 1233 00:55:18,978 --> 00:55:21,748 IT'S A PREDICTION AROUND THAT 1234 00:55:21,748 --> 00:55:22,315 DISTRIBUTION AND WHERE IT'S 1235 00:55:22,315 --> 00:55:23,883 HEADING. 1236 00:55:23,883 --> 00:55:26,286 SO JUST A COUPLE SNAP SHOTS. 1237 00:55:26,286 --> 00:55:33,226 THIS IS THE PREDICTIVE MODEL FOR 1238 00:55:33,226 --> 00:55:35,428 A TYPICAL EXPERIMENTAL MODEL FOR 1239 00:55:35,428 --> 00:55:38,898 QT PROLONGATION THE 1240 00:55:38,898 --> 00:55:39,632 PHYSIOLOGICAL OBSERVABLE WE WANT 1241 00:55:39,632 --> 00:55:39,966 TO UNDERSTAND. 1242 00:55:39,966 --> 00:55:42,602 YOU CAN SEE IT HAS SOME SIGNAL 1243 00:55:42,602 --> 00:55:43,136 THERE BUT IT IS FAR FROM 1244 00:55:43,136 --> 00:55:46,973 PERFECT. 1245 00:55:46,973 --> 00:55:51,811 SOME OF OUR MODELS ARE EVEN 1246 00:55:51,811 --> 00:55:52,112 WORSE. 1247 00:55:52,112 --> 00:55:54,714 THERE'S SORT OF A SIGNAL BUT A 1248 00:55:54,714 --> 00:55:57,417 WHOLE LOT OF MOLECULE PERFORMING 1249 00:55:57,417 --> 00:56:00,053 NOT PARTICULARLY WELL. 1250 00:56:00,053 --> 00:56:03,823 SOME ARE AT THE BOTTOM OF THE 1251 00:56:03,823 --> 00:56:05,658 ASSAY MEASURABILITY WHICH MAY 1252 00:56:05,658 --> 00:56:07,894 EXPLAIN IT BUT NONETHELESS IT IS 1253 00:56:07,894 --> 00:56:09,696 CLEAR IT'S NOT A LOUD SIGNAL 1254 00:56:09,696 --> 00:56:13,767 SHOWING UP HERE AND THE MODELS 1255 00:56:13,767 --> 00:56:16,669 ARE FAR FROM PERFECT AND SHOULD 1256 00:56:16,669 --> 00:56:18,905 NOT USE THEM FROM SIGNAL POINT 1257 00:56:18,905 --> 00:56:20,507 PREDICTION BUT ENOUGH SIGNAL TO 1258 00:56:20,507 --> 00:56:21,407 MOVE THAT DISTRIBUTION IN THE 1259 00:56:21,407 --> 00:56:24,444 DIRECTION WE WANT IT TO GO. 1260 00:56:24,444 --> 00:56:34,988 IT HIGHLIGHTS THE ESSENTIALITY 1261 00:56:36,089 --> 00:56:38,558 LOOP OF WET LAB AND DRY LAB IN 1262 00:56:38,558 --> 00:56:39,125 THE PROCESS. 1263 00:56:39,125 --> 00:56:43,263 JUST A THOUGHT THAT OCCURS TO ME 1264 00:56:43,263 --> 00:56:45,064 FROM THE TIME I'VE WATCHED THIS 1265 00:56:45,064 --> 00:56:47,267 WORK GOING ON, I THINK IT IS 1266 00:56:47,267 --> 00:56:49,102 HELPFUL IF WE THINK ABOUT SOME 1267 00:56:49,102 --> 00:56:51,237 OF AS THE GENERATIVE DESIGN 1268 00:56:51,237 --> 00:56:53,740 ALGORITHMS BY THINKING ABOUT THE 1269 00:56:53,740 --> 00:56:55,508 LANGUAGE THAT IS APPROPRIATE FOR 1270 00:56:55,508 --> 00:56:57,110 THE OBJECT WE'RE DESIGNING AND 1271 00:56:57,110 --> 00:57:00,446 THE TASK WE'RE DESIGNING FOR. 1272 00:57:00,446 --> 00:57:01,748 SO LET ME TELL YOU ABOUT MORE 1273 00:57:01,748 --> 00:57:05,251 ABOUT WHAT I MEAN ABOUT THAT. 1274 00:57:05,251 --> 00:57:06,619 MANY PEOPLE IN THIS VIRTUAL ROOM 1275 00:57:06,619 --> 00:57:08,454 WILL BE FAR MORE EXPERT ON THIS 1276 00:57:08,454 --> 00:57:09,422 THAN I. 1277 00:57:09,422 --> 00:57:12,325 SO I'M GOING TO JUST GIVE YOU A 1278 00:57:12,325 --> 00:57:13,359 SOMEWHAT HIGH LEVEL DESCRIPTION. 1279 00:57:13,359 --> 00:57:18,865 BUT IN THE EARLY DAYS, WE WERE 1280 00:57:18,865 --> 00:57:26,339 VERY MUCH USING VARIATIONAL AUTO 1281 00:57:26,339 --> 00:57:28,508 ENCODERS FOR DESIGN AND SPACE. 1282 00:57:28,508 --> 00:57:30,577 WE WOULD TAKE OUR MODEL AND 1283 00:57:30,577 --> 00:57:32,612 ENCODE IT INTO THAT CONTINUOUS 1284 00:57:32,612 --> 00:57:34,380 SPACE AND THEN DECODE IT AT THE 1285 00:57:34,380 --> 00:57:34,781 BACK END. 1286 00:57:34,781 --> 00:57:39,452 I WAS VERY LAZY WHEN I WAS 1287 00:57:39,452 --> 00:57:41,154 GENERATING THIS LOOP OF GRAPHIC 1288 00:57:41,154 --> 00:57:42,856 JUST USING THE SAME MOLECULE 1289 00:57:42,856 --> 00:57:45,758 INPUT AND OUTPUT. 1290 00:57:45,758 --> 00:57:47,327 BUT GENERALLY YOU'D GET MUCH 1291 00:57:47,327 --> 00:57:50,029 DIFFERENT MOLECULES AS YOU 1292 00:57:50,029 --> 00:57:53,733 NAVIGATED WITH SPACE. 1293 00:57:53,733 --> 00:58:04,277 AND OFTEN TIMES BECAUSE -- LET 1294 00:58:05,979 --> 00:58:08,281 ME STEP BACK AND SAY THIS AGAIN. 1295 00:58:08,281 --> 00:58:10,583 MOLECULES CAN BE THOUGHT OF AS 1296 00:58:10,583 --> 00:58:10,884 CONTINUOUS. 1297 00:58:10,884 --> 00:58:12,986 THEY'RE CLEARLY CONTINUOUS 1298 00:58:12,986 --> 00:58:15,188 DISTRIBUTION OF ELECTRON DENSITY 1299 00:58:15,188 --> 00:58:17,757 ACROSS A SPACE THAT EXISTS IN 1300 00:58:17,757 --> 00:58:18,324 THE WORLD. 1301 00:58:18,324 --> 00:58:23,796 BUT THEY ACT AS IF THEY WERE 1302 00:58:23,796 --> 00:58:25,765 ATOMS CONNECTED WITH BOND ANGLES 1303 00:58:25,765 --> 00:58:29,769 AND ROTATIONS. 1304 00:58:29,769 --> 00:58:32,205 A CONTINUOUS REPRESENTATION OF 1305 00:58:32,205 --> 00:58:35,308 THIS ORGANIC CHEMISTRY VIEW OF A 1306 00:58:35,308 --> 00:58:38,177 SMALL MOLECULE IS A MISMATCH IN 1307 00:58:38,177 --> 00:58:41,748 LANGUAGE STYLE AND LEADS TO US 1308 00:58:41,748 --> 00:58:45,285 GENERATING MOLECULES THAT ARE 1309 00:58:45,285 --> 00:58:46,753 SOMETIMES UGLY THAT CHEMISTS 1310 00:58:46,753 --> 00:58:48,321 CAN'T MAKE. 1311 00:58:48,321 --> 00:58:49,088 SO WE SPEND A LOT OF ENERGY 1312 00:58:49,088 --> 00:58:52,191 NEEDING TO FILTER AND APPLY 1313 00:58:52,191 --> 00:58:57,497 OTHER APPROACHES TO FORCE OUR 1314 00:58:57,497 --> 00:59:00,466 MOLECULES THAT WE DESIGN INTO 1315 00:59:00,466 --> 00:59:07,106 REASONABLE CHEMICAL SPACES. 1316 00:59:07,106 --> 00:59:11,878 CALLING THAT OUT HERE THAT WE'VE 1317 00:59:11,878 --> 00:59:13,980 DRAWN IT AS CHEMICALLY 1318 00:59:13,980 --> 00:59:15,782 REASONABLE BUT IT'S NOT ALWAYS 1319 00:59:15,782 --> 00:59:18,284 HOW THE ALGORITHMS MAKE BUT THE 1320 00:59:18,284 --> 00:59:23,890 MODELS FOR HUMAN LANGUAGES 1321 00:59:23,890 --> 00:59:24,590 ALTERS AN ALTERNATIVE PATH 1322 00:59:24,590 --> 00:59:26,626 FORWARD AND SO WE HAVE A 1323 00:59:26,626 --> 00:59:28,061 LANGUAGE FOR DESCRIBING SMALL 1324 00:59:28,061 --> 00:59:28,361 MOLECULES. 1325 00:59:28,361 --> 00:59:31,064 WE HAVE MORE THAN ONE BUT THIS 1326 00:59:31,064 --> 00:59:35,735 IS SMILES AND THIS IS A MOLECULE 1327 00:59:35,735 --> 00:59:37,537 SIMPLE ENOUGH I COULD WRITE THIS 1328 00:59:37,537 --> 00:59:39,639 BY HAND AND UNDERSTAND IT BUT WE 1329 00:59:39,639 --> 00:59:42,642 CAN ALSO DESCRIBE MORE 1330 00:59:42,642 --> 00:59:44,177 COMPLICATED MOLECULES THAT ARE 1331 00:59:44,177 --> 00:59:45,578 LESS SIMPLE TO WRITE BY HAND 1332 00:59:45,578 --> 00:59:48,147 USING THIS LANGUAGE. 1333 00:59:48,147 --> 00:59:50,950 FOR THOSE OF YOU WHO ARE 1334 00:59:50,950 --> 00:59:53,720 ASSOCIATED WITH AMGEN, YOU MUST 1335 00:59:53,720 --> 01:00:00,059 RECOGNIZE THIS AS OUR MARKETED K 1336 01:00:00,059 --> 01:00:02,028 RAS INHIBITOR FOR CERTAIN KINDS 1337 01:00:02,028 --> 01:00:03,863 OF CANCERS. 1338 01:00:03,863 --> 01:00:06,933 BECAUSE WE HAVE THIS LANGUAGE WE 1339 01:00:06,933 --> 01:00:10,403 CAN USE WE CAN USE TECHNIQUES AS 1340 01:00:10,403 --> 01:00:13,639 A CORE DESIGN MODULE. 1341 01:00:13,639 --> 01:00:15,408 WE CAN MASK PARTS OF THE 1342 01:00:15,408 --> 01:00:17,243 STRUCTURE AND LEARN WHAT 1343 01:00:17,243 --> 01:00:21,114 TRANSFORMATIONS ARE POSSIBLE AND 1344 01:00:21,114 --> 01:00:23,016 MAKE PREDICTIONS OF A HUGE 1345 01:00:23,016 --> 01:00:25,084 NUMBER OF PROPERTIES ABOUT THOSE 1346 01:00:25,084 --> 01:00:29,756 MOLECULES AND WE CAN GENERATE 1347 01:00:29,756 --> 01:00:31,491 NEW MOLECULE DESIGNS. 1348 01:00:31,491 --> 01:00:33,760 SO THIS APPROACH OFFERS AN 1349 01:00:33,760 --> 01:00:35,595 ALTERNATIVE PATH WE EXPECT 1350 01:00:35,595 --> 01:00:38,197 YOU'LL HEAR MORE ABOUT AS WE GO 1351 01:00:38,197 --> 01:00:39,732 THROUGH THESE TWO DAYS FOR 1352 01:00:39,732 --> 01:00:42,435 THINKING ABOUT HOW WE DO 1353 01:00:42,435 --> 01:00:44,203 GENERATIVE DESIGN OF SMALL 1354 01:00:44,203 --> 01:00:45,738 MOLECULE STRUCTURES. 1355 01:00:45,738 --> 01:00:50,410 THE SAME APPROACHES ARE EQUALLY 1356 01:00:50,410 --> 01:00:52,845 APPLICABLE TO LARGE MOLECULE 1357 01:00:52,845 --> 01:00:55,148 THERAPEUTICS TO ANTIBODIES AND 1358 01:00:55,148 --> 01:00:57,750 PROTEIN THERAPEUTICS AND ENZYMES 1359 01:00:57,750 --> 01:00:59,719 AS THERAPEUTICS. 1360 01:00:59,719 --> 01:01:02,989 THEY'RE THE NATURAL LANGUAGE 1361 01:01:02,989 --> 01:01:08,327 WE'RE USING IS THE LANGUAGE OF 1362 01:01:08,327 --> 01:01:09,629 AMINO ACIDS AND WE CAN TAKE THE 1363 01:01:09,629 --> 01:01:11,998 MODELS GENERATED IN A DIFFERENT 1364 01:01:11,998 --> 01:01:14,100 SPACE AND APPLYING THEM IN A NEW 1365 01:01:14,100 --> 01:01:16,769 AND FRESH WAY TO THE QUESTIONS 1366 01:01:16,769 --> 01:01:17,904 THAT WE CARE ABOUT IN DRUG 1367 01:01:17,904 --> 01:01:22,508 DISCOVERY. 1368 01:01:22,508 --> 01:01:27,713 SO JUST A FEW CONCLUDING 1369 01:01:27,713 --> 01:01:29,582 REFLECTIONS FOR ALL OF US HERE 1370 01:01:29,582 --> 01:01:31,884 AND THEN TURN IT OVER TO 1371 01:01:31,884 --> 01:01:36,189 QUESTIONS AND COVERNVERSATIONS M 1372 01:01:36,189 --> 01:01:37,090 THE AUDIENCE. 1373 01:01:37,090 --> 01:01:39,659 FIRST THING I WANT TO SAY IS, 1374 01:01:39,659 --> 01:01:44,964 YES, THERE'S BEEN AMAZING, 1375 01:01:44,964 --> 01:01:47,033 AMAZING INNOVATIONS AND 1376 01:01:47,033 --> 01:01:50,770 SUCCESSES BASED ON USING A.I. 1377 01:01:50,770 --> 01:01:55,942 FOR IN SILICO DRUG DISCOVERY. 1378 01:01:55,942 --> 01:01:57,743 IF WE RELY ON THAT ALONE IT WILL 1379 01:01:57,743 --> 01:01:57,977 FAIL. 1380 01:01:57,977 --> 01:02:01,747 IT TAKES MORE THAN THAT TO 1381 01:02:01,747 --> 01:02:04,317 DISCOVER A DRUG THAT ACTUALLY IS 1382 01:02:04,317 --> 01:02:07,053 SAFE AND EFFICACIOUS FOR OUR 1383 01:02:07,053 --> 01:02:07,320 PATIENTS. 1384 01:02:07,320 --> 01:02:13,593 AND SO WE NEED TO THINK ABOUT A 1385 01:02:13,593 --> 01:02:17,763 WHOLISTIC VIEW ON THE 1386 01:02:17,763 --> 01:02:22,168 COMPUTATIONAL SIDE AND ANALYTICS 1387 01:02:22,168 --> 01:02:23,436 IN FRONT OF THE END USER AND 1388 01:02:23,436 --> 01:02:28,307 ALLOWS THIS EM TO -- THEM TO 1389 01:02:28,307 --> 01:02:29,775 MAKE SENSE AND ALGORITHMS TO 1390 01:02:29,775 --> 01:02:31,677 EXPOSE PATTERNS IN THE DATA THAT 1391 01:02:31,677 --> 01:02:33,212 LET THEM MAKE DECISIONS BASED ON 1392 01:02:33,212 --> 01:02:36,249 THE DATA THEY HAVE AT HAND. 1393 01:02:36,249 --> 01:02:38,751 AND OF COURSE PREDICTIVE MACHINE 1394 01:02:38,751 --> 01:02:41,587 LEARNING AND GENERATIVE 1395 01:02:41,587 --> 01:02:42,355 ARTIFICIAL INTELLIGENCE 1396 01:02:42,355 --> 01:02:46,159 ALGORITHMS HAVE AN IMPORTANT AND 1397 01:02:46,159 --> 01:02:47,793 EXCITING ROLE FOR US NOW AND IN 1398 01:02:47,793 --> 01:02:50,129 THE FUTURE AS ONE OF THE TOOLS 1399 01:02:50,129 --> 01:02:53,065 WE USE. 1400 01:02:53,065 --> 01:02:56,169 BUT WE ALSO SHOULDN'T ABANDON 1401 01:02:56,169 --> 01:02:58,471 THE PHYSICS BASED SIMULATIONS 1402 01:02:58,471 --> 01:03:01,107 WE'VE DONE SO MANY YEARS THAT 1403 01:03:01,107 --> 01:03:02,642 ANSWER QUESTIONS AT A DIFFERENT 1404 01:03:02,642 --> 01:03:05,745 LEVEL OF DETAIL THAT PROVIDE US 1405 01:03:05,745 --> 01:03:09,982 WITH MECHANISTIC UNDERSTANDING, 1406 01:03:09,982 --> 01:03:12,351 MECHANISTIC HYPOTHESES THAT LET 1407 01:03:12,351 --> 01:03:13,186 US PROBE THE SYSTEM IN A 1408 01:03:13,186 --> 01:03:19,091 DIFFERENT WAY. 1409 01:03:19,091 --> 01:03:21,727 FINALLY, THINKING ABOUT SYSTEMS 1410 01:03:21,727 --> 01:03:22,061 MODELLING. 1411 01:03:22,061 --> 01:03:24,130 WE HEARD JONI ALLOWED TO THIS 1412 01:03:24,130 --> 01:03:25,731 EARLIER AS WELL. 1413 01:03:25,731 --> 01:03:30,469 SINCE THE LATE '80s, EARLY '90s 1414 01:03:30,469 --> 01:03:37,743 AS AN INDUSTRY, WE MADE A BET ON 1415 01:03:37,743 --> 01:03:40,980 RATIONALE DESIGN AS THE WAY TO 1416 01:03:40,980 --> 01:03:42,582 DO DRUG DISCOVERY AND MADE THE 1417 01:03:42,582 --> 01:03:44,483 ASSUMPTION A SINGLE MOLECULE 1418 01:03:44,483 --> 01:03:45,952 INTERACTING WITH A SINGLE 1419 01:03:45,952 --> 01:03:49,088 PROTEIN WILL HAVE A BENEFICIAL 1420 01:03:49,088 --> 01:03:57,163 IMPACT ON DISEASE AND ON HEALTH. 1421 01:03:57,163 --> 01:03:58,965 WE ALL KNOW THE SITUATION IS 1422 01:03:58,965 --> 01:03:59,532 MORE COMPLICATED. 1423 01:03:59,532 --> 01:04:02,368 WE KNOW THERE ARE NETWORKS OF 1424 01:04:02,368 --> 01:04:05,738 PROTEINS AND CO-FACTORS AND 1425 01:04:05,738 --> 01:04:07,640 OTHER BIOLOGICAL FACTORS THAT 1426 01:04:07,640 --> 01:04:10,176 INTERACT IN A VERY COMPLICATED 1427 01:04:10,176 --> 01:04:13,212 NETWORK AND WE NEED TO BE 1428 01:04:13,212 --> 01:04:15,915 THINKING ABOUT NOT JUST A.I. 1429 01:04:15,915 --> 01:04:17,750 APPROACHES TO THIS BUT A HUGE 1430 01:04:17,750 --> 01:04:21,220 COLLECTION OF APPROACHES TO 1431 01:04:21,220 --> 01:04:23,089 UNDERSTANDING IF I IMPINGE ON 1432 01:04:23,089 --> 01:04:24,790 THE COMPLICATED SYSTEM IN ONE OR 1433 01:04:24,790 --> 01:04:27,460 TWO OR MORE PLACES, HOW DOES 1434 01:04:27,460 --> 01:04:28,628 THAT AFFECT PERCOLATE THROUGH 1435 01:04:28,628 --> 01:04:30,896 THE SYSTEM TO HAVE THAT IMPACT 1436 01:04:30,896 --> 01:04:41,440 ON DISEASE THAT WE WANT TO HAVE? 1437 01:04:51,250 --> 01:04:52,118 DRUG DISCOVERY IS HARD. 1438 01:04:52,118 --> 01:04:53,919 THE THINGS WE NEED TO HAVE IN A 1439 01:04:53,919 --> 01:04:56,889 SINGLE MOLECULE TO ENABLE US TO 1440 01:04:56,889 --> 01:04:59,392 HAVE A HIGH PROBABILITY THIS 1441 01:04:59,392 --> 01:05:00,960 MOLECULE WILL BE SUCCESSFUL IN 1442 01:05:00,960 --> 01:05:02,428 GETTING TO THE SITE OF ACTION 1443 01:05:02,428 --> 01:05:05,131 AND HAVING THE IMPACT THAT WE 1444 01:05:05,131 --> 01:05:10,202 NEED FOR EFFICACY, WHILE ALSO 1445 01:05:10,202 --> 01:05:13,439 BEING SAFE AND NOT HAVING OFF 1446 01:05:13,439 --> 01:05:15,675 TARGET EFFECTS OR EFFECTS 1447 01:05:15,675 --> 01:05:17,743 RELATED TO THE TARGET AND 1448 01:05:17,743 --> 01:05:21,480 KNOWING WE CAN MANUFACTURE THE 1449 01:05:21,480 --> 01:05:23,816 MOLECULE IN SUFFICIENT 1450 01:05:23,816 --> 01:05:25,418 QUANTITIES TO MAKE A DIFFERENCE 1451 01:05:25,418 --> 01:05:28,521 ALL HAVE TO BE IN THE SPECIAL 1452 01:05:28,521 --> 01:05:29,789 MOLECULES THAT TRANSLATE ONWARD 1453 01:05:29,789 --> 01:05:31,691 IN THE CLINIC AND FOR US TO GET 1454 01:05:31,691 --> 01:05:35,061 THERE AND TO SOLVE THOSE HARD 1455 01:05:35,061 --> 01:05:37,730 PROBLEMS, WE REQUIRE AN 1456 01:05:37,730 --> 01:05:40,333 INTEGRATED SET OF DRY LAB AND 1457 01:05:40,333 --> 01:05:42,435 WET LAB TOOLS TO MAKE PROGRESS. 1458 01:05:42,435 --> 01:05:44,070 FINALLY, I HOPE YOU'VE HEARD A 1459 01:05:44,070 --> 01:05:45,938 THREAT COMING THROUGH IN THE 1460 01:05:45,938 --> 01:05:49,642 REMARKS THAT AS WE CARRY OUT OUR 1461 01:05:49,642 --> 01:05:51,077 COMPUTATIONAL EXPERIMENTS, WE 1462 01:05:51,077 --> 01:05:54,947 NEED TO DO SO WITH RIGOR. 1463 01:05:54,947 --> 01:05:56,649 RIGOR IN THE WAY DATA HAS BEEN 1464 01:05:56,649 --> 01:05:57,717 COLLECTED. 1465 01:05:57,717 --> 01:06:00,920 RIGOR IN THE WAY DATA HAS BEEN 1466 01:06:00,920 --> 01:06:03,622 ORGANIZED AND CATEGORIZED AND 1467 01:06:03,622 --> 01:06:04,690 HAD THE METADATA PLACED AROUND 1468 01:06:04,690 --> 01:06:05,224 IT. 1469 01:06:05,224 --> 01:06:09,228 RIGOR IN THE WAY WE APPROACH OUR 1470 01:06:09,228 --> 01:06:13,132 COMPUTATIONS AND BEING VERY 1471 01:06:13,132 --> 01:06:14,834 CAREFUL TO NOT FOOL OURSELVES. 1472 01:06:14,834 --> 01:06:17,737 MANY HAVE HEARD ME SAY MANY 1473 01:06:17,737 --> 01:06:20,673 TIMES THAT I LOVED MOVING TO 1474 01:06:20,673 --> 01:06:21,340 COMPUTATIONAL CHEMISTRY BECAUSE 1475 01:06:21,340 --> 01:06:25,077 I NEVER GOT TAR IN MY TEST TUBE. 1476 01:06:25,077 --> 01:06:32,051 MY CULTURES NEVER DIED I NEVER 1477 01:06:32,051 --> 01:06:33,352 HAD MY EXPERIMENT -- MY 1478 01:06:33,352 --> 01:06:34,954 EXPERIMENT ALWAYS WORKED UNLESS 1479 01:06:34,954 --> 01:06:36,021 THE PROGRAM CRASHED AND THAT'S 1480 01:06:36,021 --> 01:06:38,758 BOTH A STRENGTH AND WEAKNESS. 1481 01:06:38,758 --> 01:06:41,427 I ALWAYS GET AN ANSWER BUT HOW 1482 01:06:41,427 --> 01:06:45,698 DO I KNOW IF THAT ANSWER HAS ANY 1483 01:06:45,698 --> 01:06:47,767 MEANING FOR DECISION MAKING FOR 1484 01:06:47,767 --> 01:06:49,034 DRUG DISCOVERY. 1485 01:06:49,034 --> 01:06:51,404 THAT IS THE PLACE WHERE WE HAVE 1486 01:06:51,404 --> 01:06:53,139 TO BE TRULY RIGOROUS AND NOT 1487 01:06:53,139 --> 01:06:53,506 FOOL OURSELVES. 1488 01:06:53,506 --> 01:06:56,008 AND FROM THE BEGINNING OF MY 1489 01:06:56,008 --> 01:07:00,613 TALK I LIKE TO REITERATE THAT 1490 01:07:00,613 --> 01:07:03,315 COMPUTATIONAL CHEMISTRY IS A 1491 01:07:03,315 --> 01:07:05,084 DISCIPLINE WITH ITS OWN RICH 1492 01:07:05,084 --> 01:07:08,053 HISTORY AND IT'S OWN EXCITE 1493 01:07:08,053 --> 01:07:11,056 BEING FUTURE AND THAT WOMEN AND 1494 01:07:11,056 --> 01:07:15,494 THE DIVERSE COMMUNITY OF 1495 01:07:15,494 --> 01:07:17,730 PRACTITIONERS HAVE BEEN IN THE 1496 01:07:17,730 --> 01:07:19,932 FIELD IN THE EARLIEST DAYS AND 1497 01:07:19,932 --> 01:07:20,900 WILL BE ON TO THE FUTURE AS 1498 01:07:20,900 --> 01:07:21,233 WELL. 1499 01:07:21,233 --> 01:07:22,601 THE BODY OF WORK I TALKED TO YOU 1500 01:07:22,601 --> 01:07:24,703 ABOUT TODAY CAME FROM THIS SET 1501 01:07:24,703 --> 01:07:29,141 OF ORGANIZATIONS AND WITH THAT, 1502 01:07:29,141 --> 01:07:31,677 SARINE, I THINK I DID OKAY ON 1503 01:07:31,677 --> 01:07:31,877 TIME. 1504 01:07:31,877 --> 01:07:33,546 >> PERFECT. 1505 01:07:33,546 --> 01:07:34,547 >> I AM GOING TO STOP SHARING 1506 01:07:34,547 --> 01:07:41,487 AND WE CAN GO TO A Q&A SESSION. 1507 01:07:41,487 --> 01:07:43,289 >> THANK YOU SO MUCH, MARTI. 1508 01:07:43,289 --> 01:07:45,758 SO INSPIRING TO HEAR ABOUT ALL 1509 01:07:45,758 --> 01:07:50,296 YOUR REFLECTIONS IS AMAZING AND 1510 01:07:50,296 --> 01:07:52,965 TO HEAR A LITTLE BIT ABOUT ASSAY 1511 01:07:52,965 --> 01:07:55,701 AS WELL BECAUSE I WAS AN ASSAY 1512 01:07:55,701 --> 01:07:57,069 BIOLOGIST AND WHAT A WONDERFUL 1513 01:07:57,069 --> 01:07:58,537 COLLABORATION AND YOU TALKING 1514 01:07:58,537 --> 01:08:00,606 ABOUT WOMEN BEING PART OF THIS 1515 01:08:00,606 --> 01:08:02,675 IS JUST EXTREMELY INSPIRING. 1516 01:08:02,675 --> 01:08:03,909 THANK YOU SO MUCH FOR DOING 1517 01:08:03,909 --> 01:08:04,109 THAT. 1518 01:08:04,109 --> 01:08:06,445 LOTS OF QUESTIONS, MARTI. 1519 01:08:06,445 --> 01:08:07,980 WE'LL START TACKLING SOME OF 1520 01:08:07,980 --> 01:08:08,347 THEM. 1521 01:08:08,347 --> 01:08:09,014 THANK YOU SO MUCH FOR EVERYONE 1522 01:08:09,014 --> 01:08:10,216 THAT ASKED QUESTIONS. 1523 01:08:10,216 --> 01:08:12,918 WE'RE GOING TO TACKLE AS MANY AS 1524 01:08:12,918 --> 01:08:14,220 WE CAN TODAY. 1525 01:08:14,220 --> 01:08:18,824 SO FIRST QUESTION WE HAVE HERE 1526 01:08:18,824 --> 01:08:21,327 HAS ANY STUDY COMPREHENSIVELY 1527 01:08:21,327 --> 01:08:23,762 COMPARED THE EXPERIMENTAL 1528 01:08:23,762 --> 01:08:25,731 PROTEIN STRUCTURES IN THE PDB TO 1529 01:08:25,731 --> 01:08:27,333 THEIR ALPHA FOLD MODELS? 1530 01:08:27,333 --> 01:08:29,168 >> I FEEL CERTAIN THAT WORK HAS 1531 01:08:29,168 --> 01:08:33,005 BEEN DONE BUT I CAN'T POINT TO 1532 01:08:33,005 --> 01:08:35,040 ANY SPECIFIC EXAMPLE. 1533 01:08:35,040 --> 01:08:39,245 AS WE HAVE STEVEN COME BEING 1534 01:08:39,245 --> 01:08:42,448 LATER IN THE SERIES, I WOULD 1535 01:08:42,448 --> 01:08:43,516 DEFINITELY WANT TO BE ASKING HIM 1536 01:08:43,516 --> 01:08:44,116 THAT QUESTION. 1537 01:08:44,116 --> 01:08:45,985 I WILL SAY THAT IN SOME OF 1538 01:08:45,985 --> 01:08:48,454 THE -- I DON'T ACTUALLY WORK FOR 1539 01:08:48,454 --> 01:08:49,655 A LIVING SO MUCH ANYMORE. 1540 01:08:49,655 --> 01:08:51,757 I JUST TALK TO PEOPLE FOR A 1541 01:08:51,757 --> 01:08:53,592 LIVING BUT WHEN I'VE GONE BACK 1542 01:08:53,592 --> 01:08:55,094 AND LOOKED TO STRUCTURES, 1543 01:08:55,094 --> 01:08:57,229 THEY'RE CLOSE ENOUGH FOR DOING 1544 01:08:57,229 --> 01:09:00,266 SOME WORK WITH AND THERE ARE 1545 01:09:00,266 --> 01:09:01,967 SOME QUESTIONS ESPECIALLY SINCE 1546 01:09:01,967 --> 01:09:07,540 THE EARLIEST VERSIONS OF THAT 1547 01:09:07,540 --> 01:09:08,874 ALGORITHM DIDN'T INCLUDE 1548 01:09:08,874 --> 01:09:10,876 CO-FACTORS OR LIGANDS WE KNOW. 1549 01:09:10,876 --> 01:09:14,246 FOR PROTEINS THAT DON'T HAVE 1550 01:09:14,246 --> 01:09:16,882 STRUCTURAL MODELS, THERE'S 1551 01:09:16,882 --> 01:09:18,584 CLEARLY SOME THINGS FOR US TO 1552 01:09:18,584 --> 01:09:20,819 LOOK ON THERE AND THAT'S 1553 01:09:20,819 --> 01:09:21,086 EXCITING. 1554 01:09:21,086 --> 01:09:23,022 I'M SURE THERE'S STUDIES LIKE 1555 01:09:23,022 --> 01:09:23,422 THAT. 1556 01:09:23,422 --> 01:09:25,124 IF ANYONE IN THE AUDIENCE KNOWS 1557 01:09:25,124 --> 01:09:27,059 OF SOME, FEEL FREE TO DROP THEM 1558 01:09:27,059 --> 01:09:28,294 IN THE CHAT FOR FOLKS TO PAY 1559 01:09:28,294 --> 01:09:28,661 ATTENTION TO. 1560 01:09:28,661 --> 01:09:29,762 >> THANK YOU, MARTI. 1561 01:09:29,762 --> 01:09:31,430 BECAUSE THIS IS A DISCUSSION AND 1562 01:09:31,430 --> 01:09:34,700 I SEE STEVEN ON THE LINE HERE, 1563 01:09:34,700 --> 01:09:37,102 WOULD YOU LIKE TO ADD ANYTHING 1564 01:09:37,102 --> 01:09:47,613 TO THAT IF YOU ARE LISTENING? 1565 01:09:57,122 --> 01:10:00,659 >> WHAT ARE THE FUTURE TRENDS OR 1566 01:10:00,659 --> 01:10:02,661 A.I. METHODS LIKE ALPHA FOLDS? 1567 01:10:02,661 --> 01:10:08,734 >> ALL THE ABOVE. 1568 01:10:08,734 --> 01:10:11,937 IT'S A VERY LAME ANSWER BUT 1569 01:10:11,937 --> 01:10:15,474 HAPPENS TO BE WHAT I TRULY 1570 01:10:15,474 --> 01:10:15,708 BELIEVE. 1571 01:10:15,708 --> 01:10:20,512 I DO THINK QUANTUM COMPUTING AND 1572 01:10:20,512 --> 01:10:26,518 THINKING IN KUBBITZ MAYBE OPENS 1573 01:10:26,518 --> 01:10:29,955 UP POSSIBILITIES FOR US TO THINK 1574 01:10:29,955 --> 01:10:31,090 ABOUT THE STATE FUNCTION AND A 1575 01:10:31,090 --> 01:10:40,032 RICHER WAY TO THINK ABOUT THE 1576 01:10:40,032 --> 01:10:41,734 BOLESMAN DISTRIBUTION BECAUSE 1577 01:10:41,734 --> 01:10:44,937 WE'RE THINKING IN KUBITZ. 1578 01:10:44,937 --> 01:10:47,139 I NEVER WANT TO MISS A CHANCE TO 1579 01:10:47,139 --> 01:10:49,108 SAY THERE'S ALTERNATIVES TO 1580 01:10:49,108 --> 01:10:51,010 QUANTUM COMPUTING THAT MAY BE 1581 01:10:51,010 --> 01:10:54,279 EQUALLY INTERESTING AND RICH AND 1582 01:10:54,279 --> 01:10:56,281 MAY HAVE A LOWER ENERGY 1583 01:10:56,281 --> 01:10:59,018 FOOTPRINT THAN QUANTUM 1584 01:10:59,018 --> 01:10:59,284 COMPUTING. 1585 01:10:59,284 --> 01:11:01,754 SO YES, THESE NEW METHODS OF 1586 01:11:01,754 --> 01:11:05,724 COMPUTING, ABSOLUTELY TRUE AND 1587 01:11:05,724 --> 01:11:10,596 YES, METADYNAMICS AND UNDERSTAND 1588 01:11:10,596 --> 01:11:12,264 GOING ON IN THE PROTEIN SYSTEMS 1589 01:11:12,264 --> 01:11:17,736 AND HOW THEY INTEGRATE INTO 1590 01:11:17,736 --> 01:11:19,872 NETWORKS AND I CLEARLY BELIEVE 1591 01:11:19,872 --> 01:11:21,974 THAT'S A PLACE WE SHOULD PAY 1592 01:11:21,974 --> 01:11:22,908 ATTENTION TO BUT I'D REALLY 1593 01:11:22,908 --> 01:11:28,280 CHALLENGE ALL OF YOU AS YOU 1594 01:11:28,280 --> 01:11:35,721 THINK ABOUT THINKING OF THE 1595 01:11:35,721 --> 01:11:36,822 TRANSLATIONAL NEEDS OF OUR 1596 01:11:36,822 --> 01:11:37,956 PATIENTS, WHAT ARE THIS OPEN 1597 01:11:37,956 --> 01:11:39,458 QUESTIONS OUT THERE AND HOW CAN 1598 01:11:39,458 --> 01:11:45,097 WE THINK OF KNITTING TOGETHER 1599 01:11:45,097 --> 01:11:46,632 THE TECHNIQUE AND TACKLING NEW 1600 01:11:46,632 --> 01:11:50,769 ONES THAT TACKLE FUNDAMENTAL 1601 01:11:50,769 --> 01:11:51,036 QUESTIONS. 1602 01:11:51,036 --> 01:11:53,338 THAT'S WHERE THE NOBELS FOR 1603 01:11:53,338 --> 01:11:54,306 CHEMISTRY AND MEDICINE AND 1604 01:11:54,306 --> 01:11:56,008 PHYSICS WILL COME IN THE FUTURE 1605 01:11:56,008 --> 01:11:58,944 IS FROM THINKING ABOUT IT FROM 1606 01:11:58,944 --> 01:12:03,649 THAT ANGLE AND USING THESE AS 1607 01:12:03,649 --> 01:12:03,849 TOOLS. 1608 01:12:03,849 --> 01:12:06,185 >> THANK YOU SO MUCH, MARTI. 1609 01:12:06,185 --> 01:12:08,020 MORE QUESTIONS SO, LET'S TACKLE 1610 01:12:08,020 --> 01:12:08,921 SOME MORE. 1611 01:12:08,921 --> 01:12:14,893 HOW CAN WE IMPROVE THE 1612 01:12:14,893 --> 01:12:19,932 VALIDATION OF A.I. M.L. S. 1613 01:12:19,932 --> 01:12:22,134 >> MY FRIEND WILL TALK TO YOU 1614 01:12:22,134 --> 01:12:25,204 TOMORROW ABOUT THAT EXACT TOPIC. 1615 01:12:25,204 --> 01:12:32,077 I DO KNOW I HAVE BIASES FROM MY 1616 01:12:32,077 --> 01:12:34,880 OWN HISTORY WHERE A DEEP 1617 01:12:34,880 --> 01:12:39,451 PREFERENCE FOR FEWER MODEL 1618 01:12:39,451 --> 01:12:41,220 PARAMETERS RATHER THAN MORE 1619 01:12:41,220 --> 01:12:44,389 MODEL PARAMETERS OF AN INTENSE 1620 01:12:44,389 --> 01:12:47,292 DESIRE WE NOT OVER FIT OUR 1621 01:12:47,292 --> 01:12:48,861 MODELS JUST BECAUSE WE HAVE SO 1622 01:12:48,861 --> 01:12:54,366 MUCH SPACE AND SO MUCH ROPE TO 1623 01:12:54,366 --> 01:13:01,106 HANG OURSELVES AND WOULD HEAR A 1624 01:13:01,106 --> 01:13:04,510 THEME AROUND THE NEED FOR 1625 01:13:04,510 --> 01:13:05,911 PROSPECTIVE, BLIND CHALLENGES 1626 01:13:05,911 --> 01:13:07,980 WHERE WE DON'T ALREADY KNOW THE 1627 01:13:07,980 --> 01:13:10,315 ANSWER AND MAKE BOLD AND 1628 01:13:10,315 --> 01:13:13,118 PREDICTIVE PREDICTIONS WE THEN 1629 01:13:13,118 --> 01:13:16,722 GO BACK AND TEST AND THEN 1630 01:13:16,722 --> 01:13:20,058 FINALLY ALWAYS HAVING VERY 1631 01:13:20,058 --> 01:13:22,995 CAREFULLY AND WELL CHOSEN NUL 1632 01:13:22,995 --> 01:13:25,097 MODELS TO WHICH WE COMPARE. 1633 01:13:25,097 --> 01:13:29,768 SO IT IS NOT ENOUGH TO SAY MY 1634 01:13:29,768 --> 01:13:35,274 NEW METHOD DOES GREAT ON THE 1635 01:13:35,274 --> 01:13:35,507 SYSTEMS. 1636 01:13:35,507 --> 01:13:39,344 IT NEEDS TO HAVE UNDERSTANDING 1637 01:13:39,344 --> 01:13:41,547 AND NEEDS TO COMPARE TO OTHER 1638 01:13:41,547 --> 01:13:43,982 MUCH SIMPLER MODELS SO WE KNOW 1639 01:13:43,982 --> 01:13:49,721 THE TRUE ADDED VALUE OF THE TIME 1640 01:13:49,721 --> 01:13:51,290 AND ENERGY AND INTELLECT WE'RE 1641 01:13:51,290 --> 01:13:52,191 SPENDING ON THAT MODEL. 1642 01:13:52,191 --> 01:13:57,362 BUT I AM GOING TO STOP THERE 1643 01:13:57,362 --> 01:14:00,666 BECAUSE I KNOW THAT THERE'LL BE 1644 01:14:00,666 --> 01:14:02,167 INTERESTING AND USEFUL STUFF 1645 01:14:02,167 --> 01:14:02,968 WITH THAT TOMORROW. 1646 01:14:02,968 --> 01:14:03,802 >> THANK YOU SO MUCH. 1647 01:14:03,802 --> 01:14:04,269 THAT WAS VERY HELPFUL. 1648 01:14:04,269 --> 01:14:07,940 THANK YOU. 1649 01:14:07,940 --> 01:14:11,243 ANOTHER QUESTION HERE, HOW DOES 1650 01:14:11,243 --> 01:14:13,745 THE CONCEPT OF GENERATIVE 1651 01:14:13,745 --> 01:14:16,215 MOLECULAR DESIGN IS DIFFERENT 1652 01:14:16,215 --> 01:14:22,788 FROM SCAFFOLD OR ATOM HOPPING? 1653 01:14:22,788 --> 01:14:24,323 >> HMM. 1654 01:14:24,323 --> 01:14:25,457 INTERESTING QUESTION. 1655 01:14:25,457 --> 01:14:30,562 IN SOME WAYS IT'S NOT REALLY 1656 01:14:30,562 --> 01:14:34,032 THAT DIFFERENT. 1657 01:14:34,032 --> 01:14:36,301 JUST LIKE I SAID WE'VE BEEN 1658 01:14:36,301 --> 01:14:37,202 DOING MACHINE LEARNING SINCE 1659 01:14:37,202 --> 01:14:39,371 BEFORE THERE WERE MACHINES TO 1660 01:14:39,371 --> 01:14:40,305 LEARN ON. 1661 01:14:40,305 --> 01:14:42,608 WE PROBABLY HAVE BEEN DOING 1662 01:14:42,608 --> 01:14:44,142 GENERATIVE DESIGN SINCE BEFORE 1663 01:14:44,142 --> 01:14:45,811 THERE WERE GENERATIVE A.I. 1664 01:14:45,811 --> 01:14:46,211 ALGORITHMS. 1665 01:14:46,211 --> 01:14:51,283 SO IN SOME WAYS IT'S A NEW NAME 1666 01:14:51,283 --> 01:14:55,087 FOR AN OLD CONCEPT. 1667 01:14:55,087 --> 01:14:57,756 SOME OF THE NEW ALGORITHMIC 1668 01:14:57,756 --> 01:15:06,265 THINGS ARE NOVE -- INNOVATIONS 1669 01:15:06,265 --> 01:15:10,869 BUT WHEN I WAS USING FLOW, QXP, 1670 01:15:10,869 --> 01:15:15,207 ALGORITHM IN THE OLDEST WAYS 1671 01:15:15,207 --> 01:15:17,109 YOU'VE SEEN, WE HAVE THE ABILITY 1672 01:15:17,109 --> 01:15:19,578 TO IDENTIFY CONNECTION POINTS 1673 01:15:19,578 --> 01:15:25,117 AND TAKE A SET OF REAGENTS AND 1674 01:15:25,117 --> 01:15:28,153 DO THE COMPUTATIONAL SYNTHETIC 1675 01:15:28,153 --> 01:15:29,755 CHEMISTRY TO EXPLORE WHAT 1676 01:15:29,755 --> 01:15:33,659 HAPPENS AS TO THE FINDING AND 1677 01:15:33,659 --> 01:15:36,428 NOT GREAT PREDICTIVE BINDING 1678 01:15:36,428 --> 01:15:37,729 AFFINITY AS YOU MAKE THOSE PAIR 1679 01:15:37,729 --> 01:15:39,398 WISE CHANGES. 1680 01:15:39,398 --> 01:15:41,199 SOME OF WHAT'S DIFFERENT IS THE 1681 01:15:41,199 --> 01:15:42,467 INCREASE IN THE AMOUNT OF 1682 01:15:42,467 --> 01:15:45,737 COMPUTE WE HAVE AVAILABLE AND 1683 01:15:45,737 --> 01:15:48,807 THE APPROACHES FOR DOING THE 1684 01:15:48,807 --> 01:15:50,809 MULTI-OBJECTIVE OPTIMIZATION. 1685 01:15:50,809 --> 01:15:52,878 THOSE TOO HAVE A PRECEDENT IN 1686 01:15:52,878 --> 01:15:54,246 THE PAST WORK WE'VE DONE. 1687 01:15:54,246 --> 01:15:56,915 AND SO THIS IS THE PLACE WHERE I 1688 01:15:56,915 --> 01:16:00,018 STOP AND I SAY AGAIN, THERE IS 1689 01:16:00,018 --> 01:16:02,454 NOTHING NEW UNDER THE SUN BUT 1690 01:16:02,454 --> 01:16:05,023 THERE'S ALWAYS INNOVATION AND 1691 01:16:05,023 --> 01:16:05,624 GROWTH AND CREATIVITY FOR THE 1692 01:16:05,624 --> 01:16:09,594 FUTURE. 1693 01:16:09,594 --> 01:16:11,229 >> THANK YOU SO MUCH, MARTI. 1694 01:16:11,229 --> 01:16:13,031 WE'LL TAKE ONE MORE QUESTION 1695 01:16:13,031 --> 01:16:14,466 BECAUSE THERE'S SO MANY. 1696 01:16:14,466 --> 01:16:15,901 I DON'T WANT TO BLEED INTO THE 1697 01:16:15,901 --> 01:16:17,736 LUNCH BUT WANT TO TAKE ONE MORE 1698 01:16:17,736 --> 01:16:18,070 QUESTION. 1699 01:16:18,070 --> 01:16:20,105 >> I WILL SAY I'LL STAY ON AND 1700 01:16:20,105 --> 01:16:21,940 AT LEAST TYPE IN A FEW ANSWERS 1701 01:16:21,940 --> 01:16:25,110 TO THE QUESTIONS THAT ARE IN THE 1702 01:16:25,110 --> 01:16:27,913 CHAT BEFORE I LOG OFF AND GO 1703 01:16:27,913 --> 01:16:30,248 FIND MY LUNCH HERE. 1704 01:16:30,248 --> 01:16:31,984 >> THAT WOULD BE SUPER NICE OF 1705 01:16:31,984 --> 01:16:32,250 YOU. 1706 01:16:32,250 --> 01:16:34,486 WE APPRECIATE IT. 1707 01:16:34,486 --> 01:16:35,620 HERE'S ONE MORE QUESTION TO 1708 01:16:35,620 --> 01:16:42,327 ANSWER WHY AND PAMARTI WILL TRYO 1709 01:16:42,327 --> 01:16:45,731 ANSWER AS MUCH AS THE CAN AFTER. 1710 01:16:45,731 --> 01:16:48,266 AND WE'RE IN NEED OF MORE DATA 1711 01:16:48,266 --> 01:16:53,739 TO TRAIN IN SILICO MODELS? 1712 01:16:53,739 --> 01:16:55,107 >> BIOLOGICS. 1713 01:16:55,107 --> 01:16:58,310 BACK IN MY DAYS AT GSK, EVERY 1714 01:16:58,310 --> 01:16:59,911 MOLECULE MADE AND REGISTERED 1715 01:16:59,911 --> 01:17:02,781 WENT THROUGH A SET OF STANDARD 1716 01:17:02,781 --> 01:17:03,015 ASSAYS. 1717 01:17:03,015 --> 01:17:04,449 THOSE ASSAYS AS WE NOTED ARE 1718 01:17:04,449 --> 01:17:05,751 MODEL SYSTEMS FOR THE THING WE 1719 01:17:05,751 --> 01:17:07,085 CARE ABOUT. 1720 01:17:07,085 --> 01:17:09,621 BUT IT MEANT OVER TIME WE HAD 1721 01:17:09,621 --> 01:17:11,523 MILLIONS AND MILLIONS AND 1722 01:17:11,523 --> 01:17:14,359 MILLIONS OF DATA POINTS. 1723 01:17:14,359 --> 01:17:15,660 IN THE SAME ASSAY RUNNING ON THE 1724 01:17:15,660 --> 01:17:19,331 SAME ROBOTICS BY THE SAME 1725 01:17:19,331 --> 01:17:21,733 OPERATOR OVER MULTIPLE YEARS. 1726 01:17:21,733 --> 01:17:27,039 SO ON THE OTHER HAND, TO CREATE 1727 01:17:27,039 --> 01:17:29,708 A BIOLOGIC TAKES TIME AND 1728 01:17:29,708 --> 01:17:31,043 EXPRESSING SOMETHING IN CELLS 1729 01:17:31,043 --> 01:17:33,111 THAT CAN DIE ON YOU SO IT'S A 1730 01:17:33,111 --> 01:17:35,080 PLACE ON MY MIND. 1731 01:17:35,080 --> 01:17:38,550 ANOTHER PLACE ON MY MIND IS 1732 01:17:38,550 --> 01:17:40,619 AROUND NOW THAT I HAVE THE 1733 01:17:40,619 --> 01:17:42,821 PRIVILEGE TO WORK WITH THE 1734 01:17:42,821 --> 01:17:49,094 COMPUTATIONAL BIOLOGY TEAM IN MY 1735 01:17:49,094 --> 01:17:51,096 DEPARTMENT, THINKING ABOUT WHAT 1736 01:17:51,096 --> 01:17:54,900 ARE THE WAYS THAT LET US TEASE 1737 01:17:54,900 --> 01:18:01,106 APART BIOLOGY IN MORE 1738 01:18:01,106 --> 01:18:03,742 PHYSIOLOGICALLY RELEVANT MODEL 1739 01:18:03,742 --> 01:18:07,112 SYSTEMS RECOGNIZING THOSE TWO 1740 01:18:07,112 --> 01:18:09,347 ARE MODELS BUT THAT DATA FEELS 1741 01:18:09,347 --> 01:18:10,816 IMPORTANT TO ME AND REALLY THE 1742 01:18:10,816 --> 01:18:13,118 DATA THAT WOULD ALLOW US TO 1743 01:18:13,118 --> 01:18:16,088 START THINKING IN THIS MORE 1744 01:18:16,088 --> 01:18:17,722 SYSTEMS LEVEL MODELLING 1745 01:18:17,722 --> 01:18:19,691 PERSPECTIVE IS A PLACE THAT 1746 01:18:19,691 --> 01:18:21,860 FEELS ANOTHER PLACE THAT FEELS 1747 01:18:21,860 --> 01:18:23,428 RICH TO ME. 1748 01:18:23,428 --> 01:18:24,663 >> THANK YOU SO MUCH, MARTI, 1749 01:18:24,663 --> 01:18:27,766 THAT WAS SUPER HELPFUL. 1750 01:18:27,766 --> 01:18:30,735 SO SO, WE'RE GOING TO GO TO A 1751 01:18:30,735 --> 01:18:34,272 LUNCH BREAK AND NOW BACK AT 1752 01:18:34,272 --> 01:18:35,307 12:45. 1753 01:18:35,307 --> 01:18:36,675 MARTI, AGAIN, THANK YOU FOR ALL 1754 01:18:36,675 --> 01:18:38,076 OF US AT THE ORGANIZING 1755 01:18:38,076 --> 01:18:41,079 COMMITTEE AND ATTENDEES. 1756 01:18:41,079 --> 01:18:42,180 WHAT A WONDERFUL PRESENTATION. 1757 01:18:42,180 --> 01:18:44,316 IF YOU CAN JUST ANSWER SOME 1758 01:18:44,316 --> 01:18:46,785 QUESTIONS LIKE YOU SAID, THAT 1759 01:18:46,785 --> 01:18:48,920 WOULD BE FANTASTIC. 1760 01:18:48,920 --> 01:18:49,621 EVERYONE, I'LL SHARE MY SLIDE 1761 01:18:49,621 --> 01:18:52,858 DECK SO YOU SEE ON YOUR SCREENS 1762 01:18:52,858 --> 01:18:53,525 THE AGENDA. 1763 01:18:53,525 --> 01:18:55,227 SO WE DO HAVE A LUNCH BREAK 1764 01:18:55,227 --> 01:18:59,498 RIGHT NOW AND WE WILL BE BACK 1765 01:18:59,498 --> 01:19:04,736 EXACTLY AT 12:45 TO CONTINUE OUR 1766 01:19:04,736 --> 01:19:07,139 WORKSHOP WITH SESSION 1. 1767 01:19:07,139 --> 01:19:09,741 SUPER EXCITED. 1768 01:19:09,741 --> 01:19:10,546 BE BACK EXACTLY AT 12:45. THANK YOU SO MUCH. 1769 01:19:11,943 --> 01:19:12,344 1770 01:19:12,344 --> 01:19:12,410 1771 01:19:25,095 --> 01:19:28,865 >> THE SESSION'S OBJECTIVE IS TO 1772 01:19:28,865 --> 01:19:31,001 PROVIDE BEST TRACTS IN BUILDING 1773 01:19:31,001 --> 01:19:32,702 AND MAINTAINING DATABASES USED 1774 01:19:32,702 --> 01:19:36,506 FOR DEVELOPING ROBUST AND 1775 01:19:36,506 --> 01:19:37,941 RIGOROUS A.I.-BASED DRUG 1776 01:19:37,941 --> 01:19:41,511 DISCOVERY MODELS AND METHODS. 1777 01:19:41,511 --> 01:19:42,212 AND OUR FIRST SPEAKER FOR THE 1778 01:19:42,212 --> 01:19:49,552 SESSION I'M EXCITED TO INTRODUCE 1779 01:19:49,552 --> 01:19:58,128 DR. BARBARA ZDRAZIL AND HAS ABLE 1780 01:19:58,128 --> 01:20:00,730 ACADEMIC TRACK RECORD AND 1781 01:20:00,730 --> 01:20:01,965 INTERNATIONAL REPUTATION AND 1782 01:20:01,965 --> 01:20:08,505 WORKS AS CHEMBL COORDINATOR AT 1783 01:20:08,505 --> 01:20:14,477 THE INSTITUTE IN CAMBRIDGE, U.K. 1784 01:20:14,477 --> 01:20:16,746 AND AIDS IN DRUG DISCOVERY DATA 1785 01:20:16,746 --> 01:20:19,249 AND LOWER THE BARRIERS TO USAGE 1786 01:20:19,249 --> 01:20:22,252 GLOBALLY AND AUTOMATE DATA 1787 01:20:22,252 --> 01:20:23,653 EXTRACTION AND CURATION 1788 01:20:23,653 --> 01:20:25,522 PROCESSES WITH LOSS IN DATA 1789 01:20:25,522 --> 01:20:32,062 QUALITY AND TODAY HER TALK IS 1790 01:20:32,062 --> 01:20:33,530 ENTITLED 15 YEARS OF CHEMBL 1791 01:20:33,530 --> 01:20:36,633 CHALLENGES AND OPPORTUNITIES. 1792 01:20:36,633 --> 01:20:41,604 THE FLOOR IS YOURS AND LET YOU 1793 01:20:41,604 --> 01:20:51,915 SHARE YOUR SLIDES. 1794 01:20:56,486 --> 01:20:57,921 THANKS FOR THE KIND INTRODUCTION 1795 01:20:57,921 --> 01:21:01,491 AND OPPORTUNITY TO SPEAK HERE 1796 01:21:01,491 --> 01:21:02,492 TODAY. 1797 01:21:02,492 --> 01:21:05,996 THANKS A LOT TO THE ORGANIZERS. 1798 01:21:05,996 --> 01:21:09,199 I'D LIKE TO TALK TO YOU ABOUT 1799 01:21:09,199 --> 01:21:10,600 CHEMBL DRUG DISCOVERY DATABASE 1800 01:21:10,600 --> 01:21:11,301 HOPEFULLY MOST HAVE HEARD ABOUT 1801 01:21:11,301 --> 01:21:13,937 OR USED THE DATA. 1802 01:21:13,937 --> 01:21:17,507 AT THE OCCASION OF CHEMBL'S 15th 1803 01:21:17,507 --> 01:21:22,045 ANNIVERSARY WE RECENTLY LAD A 1804 01:21:22,045 --> 01:21:25,515 SYMPOSIUM AND AND WANTED TO 1805 01:21:25,515 --> 01:21:27,550 CHANGE PICTURES OF FORMER AND 1806 01:21:27,550 --> 01:21:28,952 CURRENT GROUP MEMBERS. 1807 01:21:28,952 --> 01:21:33,056 YOU SEE IN THE ABOVE PART OF THE 1808 01:21:33,056 --> 01:21:35,225 PRESENTATION WE HAD A NICE CAKE 1809 01:21:35,225 --> 01:21:36,326 AND PEOPLE ATTENDING. 1810 01:21:36,326 --> 01:21:38,428 THAT WAS A NICE THING TO DO. 1811 01:21:38,428 --> 01:21:41,264 AND BEFORE I START TALKING ABOUT 1812 01:21:41,264 --> 01:21:42,899 CHEMBL, I WANTED TO BRIEFLY 1813 01:21:42,899 --> 01:21:47,270 INTRODUCE OUR INSTITUTE. 1814 01:21:47,270 --> 01:21:52,675 SO I WORK AT THE EUROPEAN 1815 01:21:52,675 --> 01:21:53,510 BIOINFORMATICS INSTITUTE AND 1816 01:21:53,510 --> 01:21:58,181 THIS IS PART OF A LARGER 1817 01:21:58,181 --> 01:22:00,483 EUROPEAN INFRASTRUCTURE OF THE 1818 01:22:00,483 --> 01:22:02,018 EUROPEAN MOLECULAR LABORATORIES. 1819 01:22:02,018 --> 01:22:03,053 THERE ARE SIX DIFFERENT SITES 1820 01:22:03,053 --> 01:22:06,756 YOU CAN SEE ON THE SCREEN. 1821 01:22:06,756 --> 01:22:09,059 AND WE ARE BASICALLY FUNDED 1822 01:22:09,059 --> 01:22:12,128 LARGELY BY THE 20 MEMBER STATES 1823 01:22:12,128 --> 01:22:15,532 OF EMBL. 1824 01:22:15,532 --> 01:22:19,302 THE EBI WHERE I WORK HOST OPEN 1825 01:22:19,302 --> 01:22:20,737 RESOURCES FOR THE LIFE SCIENCES. 1826 01:22:20,737 --> 01:22:25,041 A FOCUS ON BIG DATA AND SEE THEM 1827 01:22:25,041 --> 01:22:26,776 ON THE SLIDE GROUPED BY 1828 01:22:26,776 --> 01:22:28,445 DIFFERENT TOPICS AND WE ARE 1829 01:22:28,445 --> 01:22:32,449 REALLY SO CAMPBELL IS IN THE 1830 01:22:32,449 --> 01:22:36,986 SECTOR OF SMALL MOLECULES AND WE 1831 01:22:36,986 --> 01:22:44,294 HOST THE CHEMICAL ENTITIES OF 1832 01:22:44,294 --> 01:22:49,866 BIOLOGICAL INTEREST DATABASE AND 1833 01:22:49,866 --> 01:22:57,540 THERE ARE OTHER RESOURCES AND 1834 01:22:57,540 --> 01:23:04,581 DATABASE FOR INSTANCE AND THE 1835 01:23:04,581 --> 01:23:06,249 DATA IN EUROPE AND THOUGH MOST 1836 01:23:06,249 --> 01:23:11,788 DON'T FOCUS ON SMALL MOLECULES 1837 01:23:11,788 --> 01:23:16,926 THEY REFERENCE OR INCLUDE SMALL 1838 01:23:16,926 --> 01:23:24,467 MOLECULE DATA AND THIS LINKS 1839 01:23:24,467 --> 01:23:27,837 WITH THE ACTIVITY DATA AND THE 1840 01:23:27,837 --> 01:23:29,506 STRUCTURAL DATABASE DATA. 1841 01:23:29,506 --> 01:23:33,042 WATCH OUT FOR THAT. 1842 01:23:33,042 --> 01:23:35,912 AND OUR CAMPUS IS GROWING SO 1843 01:23:35,912 --> 01:23:37,680 THERE'S A FOCUS TO INCREASE AND 1844 01:23:37,680 --> 01:23:44,087 EXPAND THE BIO INDUSTRIES AND IT 1845 01:23:44,087 --> 01:23:53,196 WILL BE ONE OF THE HIGHEST IN 1846 01:23:53,196 --> 01:23:54,497 COMPUTATIONAL BIOLOGY. 1847 01:23:54,497 --> 01:23:58,768 KEM BALL IS A DATABASE WHICH IS 1848 01:23:58,768 --> 01:24:00,570 NOW QUITE REACHED ITS 15 YEARS 1849 01:24:00,570 --> 01:24:02,472 THIS OCTOBER. 1850 01:24:02,472 --> 01:24:07,310 WHAT IS INTERESTING TO NOTE IS 1851 01:24:07,310 --> 01:24:14,117 OVER TIME CHAM -- CHEMBL HAS 1852 01:24:14,117 --> 01:24:18,188 GROWN IN SIZE AND IN TERMS OF 1853 01:24:18,188 --> 01:24:21,124 ITS CONTENT AND ALSO IN TERMS OF 1854 01:24:21,124 --> 01:24:23,393 THE COMPLEXITY OF THE DATA. 1855 01:24:23,393 --> 01:24:33,036 AND HAS AN EFFECT ON THE DATA 1856 01:24:33,036 --> 01:24:36,339 SCHEMA. 1857 01:24:36,339 --> 01:24:41,511 WHEN IT WAS EXTRACTED FROM THE 1858 01:24:41,511 --> 01:24:44,781 LITERATURE, STRUCTURE DATA AND 1859 01:24:44,781 --> 01:24:45,949 MOLECULES ALONG WITH THE BIO 1860 01:24:45,949 --> 01:24:47,417 ACTIVITY DATA AND ASSAY DETAILS. 1861 01:24:47,417 --> 01:24:52,488 YOU CAN SEE HERE IN THE MIDDLE 1862 01:24:52,488 --> 01:24:57,093 OF THE SLIDE IT WAS A TINY 1863 01:24:57,093 --> 01:24:59,929 DATABASE SCHEMA AND WE EXPOSE 1864 01:24:59,929 --> 01:25:01,264 TABLE TO THE PUBLIC. 1865 01:25:01,264 --> 01:25:03,933 THAT OF COURSE MEANS WE NEED TO 1866 01:25:03,933 --> 01:25:05,935 CURATE THE DATA AND THE PROCESS 1867 01:25:05,935 --> 01:25:09,038 TAKES A BIT LONGER. 1868 01:25:09,038 --> 01:25:12,108 THIS IS MORE TEDIOUS BUT HOPE TO 1869 01:25:12,108 --> 01:25:13,509 KEEP THE GOOD QUALITY THAT 1870 01:25:13,509 --> 01:25:15,044 CHEMBL IS KNOWN FOR. 1871 01:25:15,044 --> 01:25:18,047 AND WHAT YOU CAN ALSO SEE ON THE 1872 01:25:18,047 --> 01:25:20,416 SLIDE IS HOW THE DATA CONTENT 1873 01:25:20,416 --> 01:25:22,719 HAS BASICALLY CHANGED. 1874 01:25:22,719 --> 01:25:25,521 ON THE GRAPH IN THE BOTTOM LEFT 1875 01:25:25,521 --> 01:25:28,591 YOU CAN SEE HOW THE SHARE OF BIO 1876 01:25:28,591 --> 01:25:31,628 ACTIVITIES FROM THE DIFFERENT 1877 01:25:31,628 --> 01:25:34,163 MAIN SOURCES IN CHEMBL VERSUS 1878 01:25:34,163 --> 01:25:39,669 THE NATIVE DATA SETS HAS CHANGED 1879 01:25:39,669 --> 01:25:43,273 OVER TIME. 1880 01:25:43,273 --> 01:25:46,843 AND MOST IS FROM THE ADDITIONAL 1881 01:25:46,843 --> 01:25:50,079 CHEMISTRY LITERATURE BUT DONATED 1882 01:25:50,079 --> 01:25:51,681 BY DATA DONATION FROM DEDICATED 1883 01:25:51,681 --> 01:25:52,949 PROJECTS AND SO ON. 1884 01:25:52,949 --> 01:25:58,755 OF COURSE THIS CHANGES THE WAY 1885 01:25:58,755 --> 01:26:02,792 WE NEED TO THINK AND ADJUST THE 1886 01:26:02,792 --> 01:26:05,061 DATA DEPOSITION PROCESS AND THE 1887 01:26:05,061 --> 01:26:07,497 DATA IS NOW MORE DIVERSE AND 1888 01:26:07,497 --> 01:26:09,065 MORE COMPLEX WHICH INFLUENCES 1889 01:26:09,065 --> 01:26:10,300 OUR CURATION STRATEGY. 1890 01:26:10,300 --> 01:26:12,468 ON THE LEFT-HAND SIDE YOU CAN 1891 01:26:12,468 --> 01:26:14,270 SEE ALL THE DIFFERENT SOURCES SO 1892 01:26:14,270 --> 01:26:16,773 THE MAIN HIGH LEVEL SOURCES OF 1893 01:26:16,773 --> 01:26:20,643 INFORMATION THE DATA HAS BEEN 1894 01:26:20,643 --> 01:26:22,912 DEPOSITED INTO CHEMBL AND 1895 01:26:22,912 --> 01:26:27,050 LITERATURE AND DEPOSITED DATA 1896 01:26:27,050 --> 01:26:27,617 SETS. 1897 01:26:27,617 --> 01:26:30,553 THERE'S BIO ACTIVITY DATA 1898 01:26:30,553 --> 01:26:32,488 EXPECTED AND WE USED OUR 1899 01:26:32,488 --> 01:26:34,891 DATABASE FOR THAT BUT THIS IS 1900 01:26:34,891 --> 01:26:37,226 RATHER SMALL COMPARED TO THE 1901 01:26:37,226 --> 01:26:41,497 WHOLE SIZE AROUND 186,000 1902 01:26:41,497 --> 01:26:42,031 COMPOUNDS. 1903 01:26:42,031 --> 01:26:44,767 THE WAY WE IDENTIFIED THE PAPERS 1904 01:26:44,767 --> 01:26:48,604 WE WANTED TO INTEGRATE DATA FROM 1905 01:26:48,604 --> 01:26:52,475 WITH THE BIO ACTIVITY WAS 1906 01:26:52,475 --> 01:26:53,109 IDENTIFIED WITH ILLUMINATING THE 1907 01:26:53,109 --> 01:26:54,744 DRUGGABLE GENOME PROJECT. 1908 01:26:54,744 --> 01:26:57,046 IT'S A FOCUS ON UNDER STUDIED 1909 01:26:57,046 --> 01:27:04,053 TARGETS. 1910 01:27:04,053 --> 01:27:07,757 AND SOME OF THE DATA HAS MADE IT 1911 01:27:07,757 --> 01:27:10,259 INTO CHEL BALL. 1912 01:27:10,259 --> 01:27:13,629 WE HAD TWICE A DEPOSITION FROM 1913 01:27:13,629 --> 01:27:15,031 PUB CHEM. 1914 01:27:15,031 --> 01:27:16,766 THIS ISN'T SOMETHING WE'RE DOING 1915 01:27:16,766 --> 01:27:26,309 REGULARLY ON A REGULAR BASIS. 1916 01:27:26,309 --> 01:27:29,078 AND CURATE DRUG AND CANDIDATE 1917 01:27:29,078 --> 01:27:31,781 INFORMATION IN CHEMBL. 1918 01:27:31,781 --> 01:27:34,083 THIS IS NOT STORED WITH THE BIO 1919 01:27:34,083 --> 01:27:36,219 ACTIVITY DATA BUT EXTRACTED FROM 1920 01:27:36,219 --> 01:27:39,055 A VARIETY OF SOURCES THAT I'LL 1921 01:27:39,055 --> 01:27:45,495 EXPLAIN IN A MINUTE. 1922 01:27:45,495 --> 01:27:48,831 AND WHAT WE CAN SAY IS THE BIO 1923 01:27:48,831 --> 01:27:50,500 ACTIVITY DATA IN CHEMBL COVERS 1924 01:27:50,500 --> 01:27:52,235 ALL THE DIFFERENT STAGES OF THE 1925 01:27:52,235 --> 01:27:54,137 DRUG DISCOVERY AND DEVELOPMENT 1926 01:27:54,137 --> 01:27:55,037 PROCESS. 1927 01:27:55,037 --> 01:27:57,039 FROM TARGET DISCOVERY TO LEAD 1928 01:27:57,039 --> 01:28:00,009 DISCOVERY AND OPTIMIZATION AND 1929 01:28:00,009 --> 01:28:01,511 PRECLINICAL DEVELOPMENT AND THE 1930 01:28:01,511 --> 01:28:02,712 CLINICAL PHASES. 1931 01:28:02,712 --> 01:28:05,915 IN TERMS OF QUANTITY, MOST THE 1932 01:28:05,915 --> 01:28:09,252 DATA IS COVERING THE PRECLINICAL 1933 01:28:09,252 --> 01:28:15,825 STAGE. 1934 01:28:15,825 --> 01:28:20,129 AND I WANTED TO MAKE A POINT 1935 01:28:20,129 --> 01:28:21,564 THIS IS A TEDIOUS PROCESS SO WE 1936 01:28:21,564 --> 01:28:29,505 TAKE DATA FROM A VARIETY OF 1937 01:28:29,505 --> 01:28:32,041 SOURCES AND THE CLINICAL TRIALS 1938 01:28:32,041 --> 01:28:34,277 USING APPLICATIONS FOR CLINICAL 1939 01:28:34,277 --> 01:28:35,244 COMBOUNDZ AND THEN WE SUPPLEMENT 1940 01:28:35,244 --> 01:28:37,246 THOSE WITH ADDITIONAL SOURCES 1941 01:28:37,246 --> 01:28:45,521 LIKE DAILY MED FOR INDICATIONS. 1942 01:28:45,521 --> 01:28:46,622 AND DIFFERENT ANNOTATIONS. 1943 01:28:46,622 --> 01:28:51,060 ON ONE HAND WE HAVE TO OFTEN 1944 01:28:51,060 --> 01:28:54,030 NAME THE CHEMICAL TO A STRUCTURE 1945 01:28:54,030 --> 01:28:55,798 IN CHEMBL IN THE DATABASE. 1946 01:28:55,798 --> 01:28:56,766 OFTEN THOSE CHEMICAL STRUCTURES 1947 01:28:56,766 --> 01:29:04,307 ARE NANNOTATED AND INDICATE THE 1948 01:29:04,307 --> 01:29:07,877 DISEASE THE DRUG WAS INTENDED 1949 01:29:07,877 --> 01:29:14,851 FOR AND WE ALSO PROVIDE 1950 01:29:14,851 --> 01:29:15,151 INFORMATION. 1951 01:29:15,151 --> 01:29:19,889 AND WHAT IS NEW SINCE THE LAST 1952 01:29:19,889 --> 01:29:23,960 RELEASE OF CHEMBL 34 WE INCLUDED 1953 01:29:23,960 --> 01:29:27,063 DRUG FROM THE UNITED MEDICINE 1954 01:29:27,063 --> 01:29:29,499 AGENCY AND FIRST THE DATA IS 1955 01:29:29,499 --> 01:29:32,768 COLLECTED UP A SEPARATE DATABASE 1956 01:29:32,768 --> 01:29:35,838 INSTANCE WE INTERNALLY CALL DRUG 1957 01:29:35,838 --> 01:29:37,373 BASE AND MAKES ITS WAY TO CHEMBL 1958 01:29:37,373 --> 01:29:39,542 ONCE A YEAR APPROXIMATELY AND 1959 01:29:39,542 --> 01:29:43,112 FROM CHEMBL WE BASICALLY DONATE 1960 01:29:43,112 --> 01:29:53,523 THE DATA TO OPEN TARGETS. 1961 01:29:53,523 --> 01:30:03,866 AND IT'S A PLATFORM. 1962 01:30:12,275 --> 01:30:13,876 SO, WE WANTED TO MOVE FORWARD IN 1963 01:30:13,876 --> 01:30:19,815 THE PAST TWO YEARS I JOINED THE 1964 01:30:19,815 --> 01:30:21,050 CHEMBL TO FLAG. 1965 01:30:21,050 --> 01:30:26,055 FIRST WE UPDATED OUR PRODUCT 1966 01:30:26,055 --> 01:30:28,624 LIKENESS CAUSE OR HAVE AN 1967 01:30:28,624 --> 01:30:30,326 ALGORITHM TO REFLECT PROJECTS 1968 01:30:30,326 --> 01:30:34,330 AND IT WAS A FRAGMENT-BASED 1969 01:30:34,330 --> 01:30:37,500 APPROACH AND NOW UPDATED THIS TO 1970 01:30:37,500 --> 01:30:40,536 THE ALGORITHM PUBLISHED SOME 1971 01:30:40,536 --> 01:30:50,913 YEARS AGO AND TOOK AN 1972 01:30:52,648 --> 01:30:56,352 IMPLEMENTATION AND WE ALSO 1973 01:30:56,352 --> 01:30:57,353 INTRODUCED A NATIONAL PRODUCTS 1974 01:30:57,353 --> 01:30:57,553 FLAG. 1975 01:30:57,553 --> 01:31:00,222 NOW WE HAVE FLAGGING ALL 1976 01:31:00,222 --> 01:31:01,591 MOLECULES THAT WOULD NET TO A 1977 01:31:01,591 --> 01:31:03,292 MOLECULE IN THE DATABASE AS A 1978 01:31:03,292 --> 01:31:08,631 NATURAL PRODUCT IN CHEMBL. 1979 01:31:08,631 --> 01:31:11,367 WHAT I THEN DID WITH THE PRODUCT 1980 01:31:11,367 --> 01:31:21,911 LIKENESS I WANTED TO ANALYZE AND 1981 01:31:32,755 --> 01:31:36,292 THERE WAS DRUGS AND SYNTHETIC 1982 01:31:36,292 --> 01:31:39,061 COMPOUNDS AND THIS IS BASICALLY 1983 01:31:39,061 --> 01:31:41,297 WHAT CHEMBL REFLECTS AND LOOKING 1984 01:31:41,297 --> 01:31:42,531 ACROSS THE CORE AND JOURNALS IN 1985 01:31:42,531 --> 01:31:45,067 THE RIGHT BOTTOM CORNER YOU SEE 1986 01:31:45,067 --> 01:31:46,569 IT'S APPROXIMATELY THE SAME IN 1987 01:31:46,569 --> 01:31:49,705 DIFFERENT JOURNALS BUT JOURNAL 1988 01:31:49,705 --> 01:31:51,507 OF NATURAL PRODUCTS ON AVERAGE 1989 01:31:51,507 --> 01:31:55,211 THE MOLECULES PUBLISHED IN THIS 1990 01:31:55,211 --> 01:31:57,580 JOURNAL HAVE A HIGHER LIKENESS 1991 01:31:57,580 --> 01:32:00,316 CORE WHICH ISN'T SURPRISING 1992 01:32:00,316 --> 01:32:00,549 EITHER. 1993 01:32:00,549 --> 01:32:04,420 AND I JUST COLLECTED ALL THE 1994 01:32:04,420 --> 01:32:06,389 DRUGS REPORTED IN CHEMBL WITH 1995 01:32:06,389 --> 01:32:08,824 THE FIRST APPROVAL DATE AND 1996 01:32:08,824 --> 01:32:09,859 PLOTTED THE LIKENESS CORE AND 1997 01:32:09,859 --> 01:32:12,395 WHAT YOU CAN SEE IS THERE IS A 1998 01:32:12,395 --> 01:32:17,533 SLIGHT DECREASE IN NATURAL 1999 01:32:17,533 --> 01:32:20,169 PRODUCT LIKENESS SINCE LIKE 20 2000 01:32:20,169 --> 01:32:20,736 OR 10 YEARS. 2001 01:32:20,736 --> 01:32:27,043 IT'S LIKELY GOING DOWN. 2002 01:32:27,043 --> 01:32:37,520 THIS LIKENESS AND WE STARTED A 2003 01:32:37,520 --> 01:32:38,954 WE'RE WORKING AT THE INSTITUTE 2004 01:32:38,954 --> 01:32:43,392 AND HE'S AN ORGANIC CHEMIST 2005 01:32:43,392 --> 01:32:45,061 INTERESTED IN NATURAL PRODUCTS 2006 01:32:45,061 --> 01:32:48,330 AND PUBLISHED ON PSEUDO NATURAL 2007 01:32:48,330 --> 01:32:49,365 PRODUCTS. 2008 01:32:49,365 --> 01:32:51,200 WHAT THIS IS IS BASICALLY 2009 01:32:51,200 --> 01:32:57,039 NATURAL PRODUCT FRAGMENTS 2010 01:32:57,039 --> 01:33:04,547 COMBINED IN NEW WAYS THAT CANNON 2011 01:33:04,547 --> 01:33:13,522 SYNTHESIZ 2012 01:33:13,522 --> 01:33:16,258 SYNTHESIZED AND AWE ANALYZED 2013 01:33:16,258 --> 01:33:18,594 NATURAL PRODUCT LIKENESS AND 2014 01:33:18,594 --> 01:33:19,528 PSEUDO NATURAL PRODUCT LIKENESS 2015 01:33:19,528 --> 01:33:20,229 IN THE DATABASE. 2016 01:33:20,229 --> 01:33:24,567 WHAT WE DID IS WE COMPARED 2017 01:33:24,567 --> 01:33:25,501 CLINICAL COMPOUNDS VERSUS 2018 01:33:25,501 --> 01:33:26,902 BACKGROUND COMPOUNDS. 2019 01:33:26,902 --> 01:33:30,039 ALL THE OTHER COMPOUNDS MARK AS 2020 01:33:30,039 --> 01:33:32,508 CLINICAL COMPOUNDS SO CLINICAL 2021 01:33:32,508 --> 01:33:34,677 CANDIDATES OF DRUGS IN CHEMBL 2022 01:33:34,677 --> 01:33:37,513 BECAUSE WE WANTED TO SEE DO WE 2023 01:33:37,513 --> 01:33:40,316 SEE AN ENRICHMENT OF PRODUCTS 2024 01:33:40,316 --> 01:33:41,283 OVER TIME. 2025 01:33:41,283 --> 01:33:44,153 IS THIS AN INCREASING POPULAR 2026 01:33:44,153 --> 01:33:44,653 CONCEPT. 2027 01:33:44,653 --> 01:33:48,657 SO DO WE SEE A NATURAL SELECTION 2028 01:33:48,657 --> 01:33:50,993 IN SMALL MOLECULE DRUG 2029 01:33:50,993 --> 01:33:51,260 DISCOVERY. 2030 01:33:51,260 --> 01:33:52,728 WE SAW THIS ENRICHMENT. 2031 01:33:52,728 --> 01:33:54,430 IF YOU WANT IT KNOW MORE ABOUT 2032 01:33:54,430 --> 01:33:56,098 THIS WE RECENTLY PUBLISHED THIS 2033 01:33:56,098 --> 01:33:58,634 SO PLEASE GO READ OUR ARTICLE. 2034 01:33:58,634 --> 01:34:01,504 WE'RE HAPPY TO ANSWER QUESTIONS 2035 01:34:01,504 --> 01:34:02,004 ABOUT THIS. 2036 01:34:02,004 --> 01:34:08,277 BUT IN THE FRAMEWORK OF THE 2037 01:34:08,277 --> 01:34:10,179 COLLABORATION WE PUBLISHED A 2038 01:34:10,179 --> 01:34:11,814 RELIABLE DATA SET FOR COMMUNITY 2039 01:34:11,814 --> 01:34:12,181 USE. 2040 01:34:12,181 --> 01:34:16,652 BECAUSE WE WORK IN THIS 2041 01:34:16,652 --> 01:34:17,753 ENVIRONMENT WHERE WE MAKE 2042 01:34:17,753 --> 01:34:18,988 EVERYTHING OPENLY AVAILABLE 2043 01:34:18,988 --> 01:34:22,224 WE'RE INTERESTED TO MAKE OUR 2044 01:34:22,224 --> 01:34:23,893 SOFTWARE AVAILABLE AND ALSO TO 2045 01:34:23,893 --> 01:34:24,894 CREATE THOSE DATA SETS FOR 2046 01:34:24,894 --> 01:34:26,462 COMMUNITY USE AND THE DATA SET 2047 01:34:26,462 --> 01:34:29,498 WE USED FOR THIS ANALYSIS IS A 2048 01:34:29,498 --> 01:34:33,068 DATA SET OF COMPOUND TARGETS 2049 01:34:33,068 --> 01:34:35,504 THAT HELPS EXPLORE DIFFERENCES 2050 01:34:35,504 --> 01:34:37,072 BETWEEN CLINICAL CANDIDATES AND 2051 01:34:37,072 --> 01:34:38,007 OTHER COMPOUNDS. 2052 01:34:38,007 --> 01:34:40,943 THE PAPER CAME OUT I THINK TWO 2053 01:34:40,943 --> 01:34:42,711 DAYS AGO AND IT DESCRIBES HOW 2054 01:34:42,711 --> 01:34:45,447 THE DATA SET HAS BEEN 2055 01:34:45,447 --> 01:34:46,949 ESTABLISHED AND WE ALSO PROVIDE 2056 01:34:46,949 --> 01:34:52,988 THE CODE ON GITHUB AND DIFFERENT 2057 01:34:52,988 --> 01:34:57,092 DATA RELEASES AND MAKE IT 2058 01:34:57,092 --> 01:34:58,561 AVAILABLE AND HOPING IT'S USEFUL 2059 01:34:58,561 --> 01:34:59,295 TO THE COMMUNITY. 2060 01:34:59,295 --> 01:35:01,497 FROM NATURAL PRODUCTS TO 2061 01:35:01,497 --> 01:35:09,071 CHEMICAL PROHIBITS I THINK IS AN 2062 01:35:09,071 --> 01:35:11,407 EMERGING TOPIC AND THEY'RE 2063 01:35:11,407 --> 01:35:13,542 HIGHLY SELECTIVE TOOL MOLECULES 2064 01:35:13,542 --> 01:35:15,077 THAT HELP YOU UNDERSTAND THE 2065 01:35:15,077 --> 01:35:19,381 BIOLOGY OF THE TARGET. 2066 01:35:19,381 --> 01:35:22,918 WE'RE STILL PART OF PROJECT 2067 01:35:22,918 --> 01:35:25,521 CALLED UB OPEN NOW COME TO ITS 2068 01:35:25,521 --> 01:35:27,690 END NEXT YEAR AND WHERE THE 2069 01:35:27,690 --> 01:35:33,062 EARTH RAM WAS ON ONE HAND TO 2070 01:35:33,062 --> 01:35:34,763 CREATOR SYNTHESIZE NEW HIGH 2071 01:35:34,763 --> 01:35:35,931 QUALITY OPEN ACCESS CHEMICAL 2072 01:35:35,931 --> 01:35:39,068 PROBES AND THEY SHOULD TARGET 2073 01:35:39,068 --> 01:35:41,203 NOVEL TARGET FAMILIES LIKE 2074 01:35:41,203 --> 01:35:43,739 CARRIERS OR LIGASES AND ALSO 2075 01:35:43,739 --> 01:35:47,209 ESTABLISH A CHEMO GENOMICS 2076 01:35:47,209 --> 01:35:47,476 LIBRARY. 2077 01:35:47,476 --> 01:35:52,948 AND CHEMBL WAS CHOSEN AS THE 2078 01:35:52,948 --> 01:35:54,250 DATABASE FOR THE DATA AND WE 2079 01:35:54,250 --> 01:35:54,783 DID. 2080 01:35:54,783 --> 01:36:00,456 THEN IN ADDITION, WE WORKED ON 2081 01:36:00,456 --> 01:36:04,226 CREATING A SPECIALIZED BESPOKE 2082 01:36:04,226 --> 01:36:06,929 RESOURCE FOR THE OPEN GATEWAY 2083 01:36:06,929 --> 01:36:09,431 AND SO THE WAY THE DATA IS 2084 01:36:09,431 --> 01:36:13,502 FLOWING FROM AN INTERNAL 2085 01:36:13,502 --> 01:36:19,174 DATABASE CALLED SCARAB AND 2086 01:36:19,174 --> 01:36:23,612 SEARCHS TO THE GATEWAY. 2087 01:36:23,612 --> 01:36:26,048 AND A PART FROM THIS WE STARTED 2088 01:36:26,048 --> 01:36:29,952 FLAGGING CHEMICAL PROBES BY 2089 01:36:29,952 --> 01:36:33,055 ANNOTATING PROBES FROM CHEMICAL 2090 01:36:33,055 --> 01:36:35,724 PROBES DOT-ORG AND ALSO HIGH 2091 01:36:35,724 --> 01:36:36,558 QUALITY PROBES. 2092 01:36:36,558 --> 01:36:39,495 THE REASON I MENTION THIS 2093 01:36:39,495 --> 01:36:42,164 PROJECT IN PARTICULAR IS BECAUSE 2094 01:36:42,164 --> 01:36:44,934 THIS PROJECT IS OR WAS AND IS 2095 01:36:44,934 --> 01:36:46,835 QUITE CHALLENGING BECAUSE OF THE 2096 01:36:46,835 --> 01:36:49,538 COMPLEXITY OF THE DATA THAT THEY 2097 01:36:49,538 --> 01:36:53,509 WANT THE US TO HOST IN CHEMBL. 2098 01:36:53,509 --> 01:36:55,611 IF YOU THINK OF CELL 2099 01:36:55,611 --> 01:36:57,079 AVAILABILITY DATA AND AT 2100 01:36:57,079 --> 01:36:58,113 DIFFERENT CONCENTRATIONS, 2101 01:36:58,113 --> 01:37:00,382 MULTIPLE TIME POINTS, THIS IS 2102 01:37:00,382 --> 01:37:05,287 QUITE COMPLICATED TO STORE IN A 2103 01:37:05,287 --> 01:37:07,156 RELATIONAL DATABASE LIKE CHEMBL 2104 01:37:07,156 --> 01:37:09,058 AND MAY HAVE KEPT VISUALIZATIONS 2105 01:37:09,058 --> 01:37:10,626 AND MAY BE STORED IN CHEMBL. 2106 01:37:10,626 --> 01:37:13,495 OUR SOLUTION WAS TO HOST THE KEY 2107 01:37:13,495 --> 01:37:17,333 DATA IN CHEMBL AND THEN HAVE 2108 01:37:17,333 --> 01:37:19,835 SOME ON THE GATEWAY YOU CAN SEE 2109 01:37:19,835 --> 01:37:23,172 ON THE RIGHT-HAND SIDE AND FOR 2110 01:37:23,172 --> 01:37:24,773 OTHER DATABASE LIKE IMAGES, WE 2111 01:37:24,773 --> 01:37:27,042 LINK OUT TO OTHER DATABASES LIKE 2112 01:37:27,042 --> 01:37:29,945 THE BIO IMAGE ARCHIVE WHICH IS 2113 01:37:29,945 --> 01:37:40,356 ALSO EBI DATA RESOURCE. 2114 01:37:45,127 --> 01:37:47,262 WE HAVE THE OPPORTUNITY ALREADY 2115 01:37:47,262 --> 01:37:50,199 A FEWER YEARS AGO WITH CHEMBL 2116 01:37:50,199 --> 01:37:52,534 24, THE INFRASTRUCTURE SO 2117 01:37:52,534 --> 01:37:53,502 DEVELOPMENTS HAVE BEEN MADE. 2118 01:37:53,502 --> 01:37:58,974 THERE HAS BEEN AN ACTIVITY 2119 01:37:58,974 --> 01:38:01,410 PROPERTY STABLE FOR INSTANCE 2120 01:38:01,410 --> 01:38:02,478 INCLUDED AND THESE ARE METADATA 2121 01:38:02,478 --> 01:38:05,114 TABLES THAT ALLOW YOU TO STORE 2122 01:38:05,114 --> 01:38:06,148 ADDITIONAL INFORMATION SO THAT 2123 01:38:06,148 --> 01:38:08,684 YOU DON'T HAVE TO SPLIT THAT IN 2124 01:38:08,684 --> 01:38:09,718 MULTIPLE ASSAYS BUT CAN GROUP 2125 01:38:09,718 --> 01:38:19,428 THE COMPLEX DATA ALL TOGETHER. 2126 01:38:19,428 --> 01:38:22,464 THIS LEADS ME TO TALK ABOUT DATA 2127 01:38:22,464 --> 01:38:24,133 CURATION. 2128 01:38:24,133 --> 01:38:28,837 AS WE STORE THIS WE NEED TO 2129 01:38:28,837 --> 01:38:29,505 CURATE. 2130 01:38:29,505 --> 01:38:31,473 THE MORE DATA WE GET THE MORE 2131 01:38:31,473 --> 01:38:32,975 CURATION WE HAVE TO DO. 2132 01:38:32,975 --> 01:38:34,043 THIS IS A SLIDE I WOULD USUALLY 2133 01:38:34,043 --> 01:38:36,912 SHOW AT THIS BEGINNING OF SUCH A 2134 01:38:36,912 --> 01:38:40,416 TALK BUT I THINK IT'S NICE TO 2135 01:38:40,416 --> 01:38:41,283 SHOW NOW BECAUSE IT SHOWS THE 2136 01:38:41,283 --> 01:38:44,286 DIFFERENT ASPECTS OF DATA 2137 01:38:44,286 --> 01:38:45,587 CURATION IN CHEMBL. 2138 01:38:45,587 --> 01:38:48,090 WE STANDARDIZE CHEMICAL 2139 01:38:48,090 --> 01:38:48,357 COMPOUNDS. 2140 01:38:48,357 --> 01:38:49,992 THIS IS ALL PUBLISHED AND YOU 2141 01:38:49,992 --> 01:38:52,795 CAN FIND THE WORK FLOW WE'RE 2142 01:38:52,795 --> 01:38:55,898 USING IN A PAPER AND GENERAL OF 2143 01:38:55,898 --> 01:38:58,167 CHEM INFORMATICS AND THE 2144 01:38:58,167 --> 01:38:59,902 STRUCTURE PIPELINE IS OPEN ON 2145 01:38:59,902 --> 01:39:00,636 GITHUB. 2146 01:39:00,636 --> 01:39:02,805 WE ALSO CURATE AND STANDARDIZE 2147 01:39:02,805 --> 01:39:07,276 THE BIO ACTIVITY DATA. 2148 01:39:07,276 --> 01:39:10,045 WE DO SOME FLAGGING AND 2149 01:39:10,045 --> 01:39:12,581 CALCULATE THE P CHEMBL VALUE. 2150 01:39:12,581 --> 01:39:16,318 BUT MOST THE TIME WE SPEND ON 2151 01:39:16,318 --> 01:39:17,953 CURATING ASSAY INFORMATION AND 2152 01:39:17,953 --> 01:39:19,822 THAT'S MOST RELEVANT TO THE 2153 01:39:19,822 --> 01:39:20,155 WORKSHOP. 2154 01:39:20,155 --> 01:39:21,056 WHAT'S IT MEAN? 2155 01:39:21,056 --> 01:39:24,193 ON ONE HAND YOU NEED TO ASSIGN 2156 01:39:24,193 --> 01:39:25,394 YOUR TARGET AND IT'S TEMPTING TO 2157 01:39:25,394 --> 01:39:27,896 THINK THE TARGET IN AN ASSAY IS 2158 01:39:27,896 --> 01:39:29,998 ALWAYS A SINGLE PROTEIN WHICH 2159 01:39:29,998 --> 01:39:31,366 CERTAINLY ISN'T TRUE. 2160 01:39:31,366 --> 01:39:35,404 IN MOST CASES IT'S NOT A SINGLE 2161 01:39:35,404 --> 01:39:35,637 PROTEIN. 2162 01:39:35,637 --> 01:39:37,439 OFTEN, IT'S EITHER A PROTEIN 2163 01:39:37,439 --> 01:39:40,075 COMPLEX OR PROTEIN FAMILY OR 2164 01:39:40,075 --> 01:39:42,611 EVEN SOMETHING NOT A PROTEIN 2165 01:39:42,611 --> 01:39:47,316 LIKE A CELL LINE OR TISSUE OR 2166 01:39:47,316 --> 01:39:47,816 WHOLE ORGANISM. 2167 01:39:47,816 --> 01:39:50,619 WE NEED TO ASSIGN THE TARGET AND 2168 01:39:50,619 --> 01:39:54,523 THEN LINK TO ALL KINDS OF 2169 01:39:54,523 --> 01:39:57,059 ONTOLOGIES AND WHAT USES UP MOST 2170 01:39:57,059 --> 01:40:01,497 OF THE TIME WE SPEND DURING A 2171 01:40:01,497 --> 01:40:05,501 RELEASE OR CURATION CYCLE. 2172 01:40:05,501 --> 01:40:15,978 AND AND WE'RE AWARE OF THE 2173 01:40:22,284 --> 01:40:24,853 ARROWS WHEN ONE USES OUR DATA OR 2174 01:40:24,853 --> 01:40:25,354 INTRODUCED BEFORE HAND. 2175 01:40:25,354 --> 01:40:28,524 ON ONE HAND IF YOU LOOK AT THE 2176 01:40:28,524 --> 01:40:31,026 FIGURE IN THE LEFT CORNER AT THE 2177 01:40:31,026 --> 01:40:33,061 BOTTOM HERE, THERE CAN BE 2178 01:40:33,061 --> 01:40:34,530 MISTAKES IN JOURNAL ARTICLES. 2179 01:40:34,530 --> 01:40:40,636 WE CAN'T DO ANYTHING ABOUT THIS. 2180 01:40:40,636 --> 01:40:44,540 THEN DURING OUR IN-HOUSE 2181 01:40:44,540 --> 01:40:46,408 CURATION LOOK AT MECHANISMS. 2182 01:40:46,408 --> 01:40:48,544 SO WE CAN'T REALLY INFLUENCE 2183 01:40:48,544 --> 01:40:54,183 WHAT IS HAPPENING BEFORE THE 2184 01:40:54,183 --> 01:41:03,058 DATA REACHES US BUT WE CAN 2185 01:41:03,058 --> 01:41:05,127 IMPROVE OUR DATA CHECKS AND 2186 01:41:05,127 --> 01:41:06,461 CLEAN UP OUR DATA. 2187 01:41:06,461 --> 01:41:10,899 THERE'S BEEN A RECENT ARTICLE BY 2188 01:41:10,899 --> 01:41:13,502 GREG LANDROM THAT DISCUSSES THIS 2189 01:41:13,502 --> 01:41:15,537 AND IT WAS HELPFUL AND WE'RE 2190 01:41:15,537 --> 01:41:17,773 WORKING ON RAISES AWARENESS OF 2191 01:41:17,773 --> 01:41:20,509 THE STRUCTURE AND ANNOTATIONS 2192 01:41:20,509 --> 01:41:24,146 AND ALSO WANT TO AT SOME POINT 2193 01:41:24,146 --> 01:41:25,881 RELEASE OUR DATA CHECK NOTEBOOK 2194 01:41:25,881 --> 01:41:26,515 TO THE COMMUNITY AND MAKE MORE 2195 01:41:26,515 --> 01:41:31,653 RECOMMENDATIONS. 2196 01:41:31,653 --> 01:41:34,122 THE REASON I MENTIONED THIS IS 2197 01:41:34,122 --> 01:41:37,059 YOU NEED TO BE AWARE IT DEPENDS 2198 01:41:37,059 --> 01:41:38,393 ON THE USE CASE HOW MUCH YOU 2199 01:41:38,393 --> 01:41:41,797 WANT TO REFINE THE DATA SET AND 2200 01:41:41,797 --> 01:41:43,432 WHAT YOU PUT OUT WHEN YOU USE 2201 01:41:43,432 --> 01:41:45,067 THE DATA FOR MODELLING PURPOSE. 2202 01:41:45,067 --> 01:41:47,402 SO HERE'S ONE EXAMPLE. 2203 01:41:47,402 --> 01:41:51,607 YOU CAN COLLECT ALL THE DATA OR 2204 01:41:51,607 --> 01:41:53,508 YOU CAN FURTHER FIELD FOR THE 2205 01:41:53,508 --> 01:41:55,210 HUMAN TARGET. 2206 01:41:55,210 --> 01:42:01,516 OR YOU CAN GO EVEN DEEPER AND 2207 01:42:01,516 --> 01:42:02,884 JUST CONSIDER WILD TYPE. 2208 01:42:02,884 --> 01:42:05,120 THESE ARE ALL QUESTIONS YOU NEED 2209 01:42:05,120 --> 01:42:06,989 TO ASK YOURSELF BEFORE USING OUR 2210 01:42:06,989 --> 01:42:07,189 DATA. 2211 01:42:07,189 --> 01:42:08,790 WE CAN'T TAKE RESPONSIBILITY FOR 2212 01:42:08,790 --> 01:42:10,792 HOW YOU ARE USING OUR DATA BUT 2213 01:42:10,792 --> 01:42:13,962 WE THINK WE ARE IN A POSITION TO 2214 01:42:13,962 --> 01:42:15,063 SHORTLY RELEASE MORE 2215 01:42:15,063 --> 01:42:16,098 RECOMMENDATIONS. 2216 01:42:16,098 --> 01:42:20,235 I DON'T HAVE TIME TO GO INTO A 2217 01:42:20,235 --> 01:42:22,738 LOT OF DETAIL HERE BUT WE DO A 2218 01:42:22,738 --> 01:42:25,307 LOT OF DATA CHECKS FOR DIFFERENT 2219 01:42:25,307 --> 01:42:26,575 ASPECTS OF THE DATA. 2220 01:42:26,575 --> 01:42:28,343 AND JUST TO END THIS TALK, I 2221 01:42:28,343 --> 01:42:30,712 WANTED TO MENTION TWO MORE 2222 01:42:30,712 --> 01:42:31,113 THINGS. 2223 01:42:31,113 --> 01:42:33,782 SO PROTEIN ANNOTATIONS IN CHEMBL 2224 01:42:33,782 --> 01:42:35,350 IS MAYBE SOMETHING NOT SO WELL 2225 01:42:35,350 --> 01:42:38,687 KNOWN SO WE HAVE BEEN 2226 01:42:38,687 --> 01:42:41,089 INTRODUCING VARYING SEQUENCES 2227 01:42:41,089 --> 01:42:42,557 TABLE IN CHEMBL 27. 2228 01:42:42,557 --> 01:42:44,192 THIS INFORMATION IS AVAILABLE 2229 01:42:44,192 --> 01:42:45,060 BUT IT'S NOT INFORMATION YOU 2230 01:42:45,060 --> 01:42:49,865 WILL FIND AT THE TARGET LEVEL. 2231 01:42:49,865 --> 01:42:53,302 BUT AT THE ASSAY LEVEL AND YOU 2232 01:42:53,302 --> 01:42:56,571 NEED TO BE AWARE WHEN YOU 2233 01:42:56,571 --> 01:42:58,307 COMBINE DATA FROM DIFFERENT 2234 01:42:58,307 --> 01:43:00,275 TARGETS AND HAVE TO CONSIDER THE 2235 01:43:00,275 --> 01:43:03,078 PROTEIN VARIANT ANNOTATIONS. 2236 01:43:03,078 --> 01:43:06,682 WE LATELY WORK TOGETHER WITH A 2237 01:43:06,682 --> 01:43:11,019 LAB TO UPDATE THE DATA VARIED 2238 01:43:11,019 --> 01:43:11,687 EXTRACTION PIPELINE. 2239 01:43:11,687 --> 01:43:14,890 SO THEY DID IT IN A SLIGHTLY 2240 01:43:14,890 --> 01:43:16,625 DIFFERENT WAY AND TOGETHER WE 2241 01:43:16,625 --> 01:43:20,629 COULD ENHANCE THESE ANNOTATIONS. 2242 01:43:20,629 --> 01:43:23,498 THIS HAS BEEN PUBLISHED IN THE 2243 01:43:23,498 --> 01:43:25,534 ARCHIVE SO HAVE A LOOK AT THE 2244 01:43:25,534 --> 01:43:27,069 WORK WE DID HERE. 2245 01:43:27,069 --> 01:43:30,138 THAT'S MY FINAL SLIDE AND FOR ME 2246 01:43:30,138 --> 01:43:31,440 THE MOST IMPORTANT ONE AND IT'S 2247 01:43:31,440 --> 01:43:36,578 UNFORTUNATE I CAN'T TALK LONGER 2248 01:43:36,578 --> 01:43:45,053 ABOUT THIS THIS IS ALSO 2249 01:43:45,053 --> 01:43:47,856 UNPUBLISHED CONTENT AND WE HAVE 2250 01:43:47,856 --> 01:43:49,124 THIS AMAZING DATA BUT HOW DO WE 2251 01:43:49,124 --> 01:43:51,793 KNOW WHAT WE CAN CAN COMBINE TO 2252 01:43:51,793 --> 01:43:53,028 BUILD MACHINE LEARNING MODELS 2253 01:43:53,028 --> 01:43:54,296 AND WE'RE TRYING TO IMPROVE 2254 01:43:54,296 --> 01:43:59,968 ASSAY AN OHIO -- ANNOTATIONS 2255 01:43:59,968 --> 01:44:02,170 AND WE LOOK AT THE END POINT AND 2256 01:44:02,170 --> 01:44:05,507 FORMAT AND THE READ OUT LIKE 2257 01:44:05,507 --> 01:44:07,409 IC50 K I. 2258 01:44:07,409 --> 01:44:08,176 WHAT WE'RE NOT DOING IS FOR 2259 01:44:08,176 --> 01:44:11,012 INSTANCE ANNOTATING THE 2260 01:44:11,012 --> 01:44:13,482 EXPERIMENTAL METHODS. 2261 01:44:13,482 --> 01:44:18,320 WE DEVELOPED MODELS TO NOW 2262 01:44:18,320 --> 01:44:20,188 IDENTIFY THE DESCRIPTION AND MAP 2263 01:44:20,188 --> 01:44:22,657 IT TO THE ONTOLOGY. 2264 01:44:22,657 --> 01:44:24,025 THIS IS WORK IN PROGRESS BUT 2265 01:44:24,025 --> 01:44:25,494 HOPING WE CAN RELEASE THIS 2266 01:44:25,494 --> 01:44:28,029 INFORMATION NEXT YEAR. 2267 01:44:28,029 --> 01:44:29,297 AND WE ALSO IN THE COURSE OF 2268 01:44:29,297 --> 01:44:31,700 WRITING A PAPER ABOUT THIS I 2269 01:44:31,700 --> 01:44:32,968 THINK MY TIME IS OVER BUT I'M 2270 01:44:32,968 --> 01:44:34,536 ALSO COMING TO AN END. 2271 01:44:34,536 --> 01:44:37,506 I WANT TO THANK THE WHOLE TEAM. 2272 01:44:37,506 --> 01:44:41,510 SO IT'S A LARGE TEAM AROUND 25 2273 01:44:41,510 --> 01:44:43,378 PEOPLE THAT ARE MAINTAINING OUR 2274 01:44:43,378 --> 01:44:44,045 FULL RESOURCES AND THESE PEOPLE 2275 01:44:44,045 --> 01:44:47,115 ARE LISTED HERE. 2276 01:44:47,115 --> 01:44:48,717 THE PEOPLE INVOLVED IN THE 2277 01:44:48,717 --> 01:44:51,052 PROJECT I'VE PRESENTED. 2278 01:44:51,052 --> 01:44:53,121 AND THEN ALSO SOME FORMER 2279 01:44:53,121 --> 01:44:54,022 MEMBERS WHO HAVE BEEN PART OF 2280 01:44:54,022 --> 01:44:57,492 THESE BIG PROJECTS AND SOME 2281 01:44:57,492 --> 01:44:59,127 MEMBERS FROM ANOTHER LAB THAT 2282 01:44:59,127 --> 01:45:01,496 WERE WORKING WITH US ON THE 2283 01:45:01,496 --> 01:45:02,097 VARIANT PIPELINE. 2284 01:45:02,097 --> 01:45:08,336 AND LAST BUT NOT LEAST I WANT TO 2285 01:45:08,336 --> 01:45:09,805 SAY THE CAKE IS ONLY HALF EATEN 2286 01:45:09,805 --> 01:45:10,672 AND THERE'S A LOT OF THINGS TO 2287 01:45:10,672 --> 01:45:17,813 DO AND LOOKING FORWARD TO THE 2288 01:45:17,813 --> 01:45:18,079 FUTURE. 2289 01:45:18,079 --> 01:45:20,081 >> THANK YOU, BARBARA, FOR A 2290 01:45:20,081 --> 01:45:20,749 WONDERFUL PRESENTATION. 2291 01:45:20,749 --> 01:45:23,885 I HEAR YOUR THOUGHTS ABOUT THE 2292 01:45:23,885 --> 01:45:26,154 IMPORTANCE OF ASSAY ANNOTATION 2293 01:45:26,154 --> 01:45:27,155 AND HOPEFULLY WE'LL BRING YOU 2294 01:45:27,155 --> 01:45:32,194 BACK TO A WEBINAR AND YOU CAN 2295 01:45:32,194 --> 01:45:34,095 TALK MORE ABOUT IT BECAUSE WE'D 2296 01:45:34,095 --> 01:45:34,863 MOVE TO HEAR THAT. 2297 01:45:34,863 --> 01:45:36,264 WE HAVE MULTIPLE QUESTIONS FOR 2298 01:45:36,264 --> 01:45:37,632 YOU SO WE'LL START TAKING 2299 01:45:37,632 --> 01:45:38,934 QUESTIONS. 2300 01:45:38,934 --> 01:45:41,503 HERE'S ONE. 2301 01:45:41,503 --> 01:45:45,507 SO HOW DO YOU ENSURE THAT 2302 01:45:45,507 --> 01:45:47,175 CAMPBELL CAPTURES SUFFICIENT 2303 01:45:47,175 --> 01:45:50,745 DETAILS AND HOW IS IT ENSURED 2304 01:45:50,745 --> 01:45:54,049 TWO IC50KI VALUES ARE REPLICATES 2305 01:45:54,049 --> 01:45:56,017 OF THE SAME EXPERIMENT FOR A 2306 01:45:56,017 --> 01:45:56,284 MOLECULE. 2307 01:45:56,284 --> 01:45:58,753 >> THE FIRST PART, HOW DO WE 2308 01:45:58,753 --> 01:46:03,058 MAKE SURE THAT WE HAVE A HOLD IN 2309 01:46:03,058 --> 01:46:04,559 THE NECESSARY CONDITIONS. 2310 01:46:04,559 --> 01:46:05,961 THE ROUGHLY NEGATIVE ANSWER IS 2311 01:46:05,961 --> 01:46:07,095 WE CAN'T. 2312 01:46:07,095 --> 01:46:09,431 WE HAVE TO BASICALLY RELY ON 2313 01:46:09,431 --> 01:46:12,133 WHAT THE DEPOSITORS WOULD GIVE 2314 01:46:12,133 --> 01:46:12,400 US. 2315 01:46:12,400 --> 01:46:14,703 WHAT WE DO HAVE IS A MANUAL THAT 2316 01:46:14,703 --> 01:46:16,671 DESCRIBES THE KIND OF 2317 01:46:16,671 --> 01:46:17,806 INFORMATION WE NEED IF DATA'S 2318 01:46:17,806 --> 01:46:22,878 BEEN DEPOSITED TO CHEMBL BUT 2319 01:46:22,878 --> 01:46:25,647 WE'RE QUITE FLEXIBLE IN TERMS 2320 01:46:25,647 --> 01:46:29,017 HOW IT'S BEING DESCRIBED IN THE 2321 01:46:29,017 --> 01:46:37,025 ASSAY DES -- DESCRIPTION AND 2322 01:46:37,025 --> 01:46:39,060 OTHERWISE YOU PRESS ANYTHING IN 2323 01:46:39,060 --> 01:46:40,495 THE SAME SCHEMA. 2324 01:46:40,495 --> 01:46:45,500 TO MAKE SURE THE NECESSARY 2325 01:46:45,500 --> 01:46:47,602 CONDITIONS ARE DELIVERED WE NEED 2326 01:46:47,602 --> 01:46:48,870 TO USE THE RIGHT BIOLOGY TO ASK 2327 01:46:48,870 --> 01:46:51,072 THE RIGHT QUESTIONS. 2328 01:46:51,072 --> 01:46:52,574 IT MAY BE SOMETIMES WE'RE 2329 01:46:52,574 --> 01:46:55,343 MISSING SOME METADATA. 2330 01:46:55,343 --> 01:46:56,978 SO WE CAN'T GUARANTEE THIS SO IN 2331 01:46:56,978 --> 01:47:00,448 THAT CASE THE DEPOSITOR TAKES 2332 01:47:00,448 --> 01:47:01,983 THE RESPONSIBILITY ON PAPERS AND 2333 01:47:01,983 --> 01:47:03,785 WE'RE RELYING ON EXTRACTORS TO 2334 01:47:03,785 --> 01:47:05,153 EX TRACT THE INFORMING. 2335 01:47:05,153 --> 01:47:06,521 WE NOW HAVE 15 YEARS OF 2336 01:47:06,521 --> 01:47:09,357 EXPERIENCE TO WORK WITH US. 2337 01:47:09,357 --> 01:47:11,493 THE USER WITH THE SAME FORMAT 2338 01:47:11,493 --> 01:47:13,261 AND STRUCTURE SO PRETTY 2339 01:47:13,261 --> 01:47:14,796 CONFIDENT THEY WOULD EXTRACT 2340 01:47:14,796 --> 01:47:17,499 EVERYTHING WE NEED BUT WE CAN'T 2341 01:47:17,499 --> 01:47:18,567 SAY WITH 100%. 2342 01:47:18,567 --> 01:47:20,569 AND THE SECOND ONE, I DON'T 2343 01:47:20,569 --> 01:47:24,873 REMEMBER NOW. 2344 01:47:24,873 --> 01:47:30,745 >> IT WAS HOW IS IT ENSURED TWO 2345 01:47:30,745 --> 01:47:33,582 OF IC50KI VALUES INTEREST THE 2346 01:47:33,582 --> 01:47:36,718 REPLICATE OF THE SAME MOLECULE? 2347 01:47:36,718 --> 01:47:38,987 >> I'M NOT SURE I UNDERSTAND. 2348 01:47:38,987 --> 01:47:45,093 USUALLY MOST THE DATA WE DON'T 2349 01:47:45,093 --> 01:47:46,995 REPORT REPLICATES AT ALL SO I'M 2350 01:47:46,995 --> 01:47:48,129 NOT SURE I UNDERSTAND THE 2351 01:47:48,129 --> 01:47:48,930 QUESTION. 2352 01:47:48,930 --> 01:47:51,066 USUALLY FOR ONE ASSAY, ONE 2353 01:47:51,066 --> 01:47:54,002 COMPOUND TARGET PAIR, GET ONE 2354 01:47:54,002 --> 01:47:54,502 MEASUREMENT. 2355 01:47:54,502 --> 01:47:57,072 IF PEOPLE WOULD WANT US TO STORE 2356 01:47:57,072 --> 01:48:01,476 REPLICATES THAT'S PART OF THE 2357 01:48:01,476 --> 01:48:01,977 METADATA WE WOULD STORE. 2358 01:48:01,977 --> 01:48:08,483 >> THANK YOU, BARBARA. 2359 01:48:08,483 --> 01:48:10,018 >> PLEASE WRITE IN THE CHAT AND 2360 01:48:10,018 --> 01:48:10,785 I CAN ANSWER LATER IF I 2361 01:48:10,785 --> 01:48:15,590 MISUNDERSTOOD THE QUESTION. 2362 01:48:15,590 --> 01:48:17,859 >> FANTASTIC. 2363 01:48:17,859 --> 01:48:20,395 ANOTHER QUESTION HERE IS IN 2364 01:48:20,395 --> 01:48:21,062 CHEMBL, DIFFERENT VARIANTS OF 2365 01:48:21,062 --> 01:48:23,965 THE SAME PROTEIN ARE ASSIGN THE 2366 01:48:23,965 --> 01:48:24,733 SAME TARGET I.D. 2367 01:48:24,733 --> 01:48:27,068 WHAT MEASURES ARE BEING TAKEN TO 2368 01:48:27,068 --> 01:48:36,811 ADDRESS THIS ISSUE? 2369 01:48:36,811 --> 01:48:38,413 WE THINK THE INFORMATION WE GIVE 2370 01:48:38,413 --> 01:48:41,483 IS NOT TO ADDRESS THE ISSUE OF 2371 01:48:41,483 --> 01:48:43,284 IDENTIFYING VARIANTS. 2372 01:48:43,284 --> 01:48:46,621 WE COULD INTRODUCE A NEW TARGET 2373 01:48:46,621 --> 01:48:48,590 IDENTIFIER FOR EVERY VARIANT. 2374 01:48:48,590 --> 01:48:50,058 THE QUESTION IS, IS IT 2375 01:48:50,058 --> 01:48:50,925 NECESSARY? 2376 01:48:50,925 --> 01:48:53,762 THE WAY WE STORE IT IS MAYBE NOT 2377 01:48:53,762 --> 01:48:58,099 SO CONVENIENT FOR EVERYONE BUT 2378 01:48:58,099 --> 01:48:59,134 THE INFORMATION IS THERE SO 2379 01:48:59,134 --> 01:49:02,237 WE'RE IN DISCUSSION TO CHANGE IT 2380 01:49:02,237 --> 01:49:02,771 OR NOT. 2381 01:49:02,771 --> 01:49:05,006 IT WOULD REQUIRE TO GO OVER 2382 01:49:05,006 --> 01:49:06,708 LEGACY DATA LIKE 14,000 TARGETS 2383 01:49:06,708 --> 01:49:10,979 TO ADDRESS THIS. 2384 01:49:10,979 --> 01:49:12,480 IT'S MORE OF AN ISSUE OF HAVING 2385 01:49:12,480 --> 01:49:13,381 THE FUNDS TO DO THAT. 2386 01:49:13,381 --> 01:49:17,152 WE MIGHT DO IT OR NOT BUT JUST 2387 01:49:17,152 --> 01:49:18,553 BE ASSURED THE INFORMATION IS 2388 01:49:18,553 --> 01:49:19,020 THERE. 2389 01:49:19,020 --> 01:49:21,289 IT'S JUST STORED AT THE ASSAY 2390 01:49:21,289 --> 01:49:21,856 LEVEL. 2391 01:49:21,856 --> 01:49:23,892 >> THANK YOU SO MUCH, BARBARA. 2392 01:49:23,892 --> 01:49:26,695 THERE'S SO MANY MORE QUESTIONS 2393 01:49:26,695 --> 01:49:29,698 SO IF YOU COULD JUST TYPE SOME 2394 01:49:29,698 --> 01:49:31,166 ANSWERS IT WILL BE FANTASTIC AND 2395 01:49:31,166 --> 01:49:32,834 WE'LL MOVE ON TO THE NEXT 2396 01:49:32,834 --> 01:49:33,068 SPEAKER. 2397 01:49:33,068 --> 01:49:34,969 WE APPRECIATE YOUR TIME. 2398 01:49:34,969 --> 01:49:35,704 THANK YOU SO MUCH. 2399 01:49:35,704 --> 01:49:36,237 ALL RIGHT. 2400 01:49:36,237 --> 01:49:40,241 LET'S ME OF ON TO THE NEXT 2401 01:49:40,241 --> 01:49:45,080 SPEAKER WHO IS DR. STEVEN 2402 01:49:45,080 --> 01:49:45,313 BURLEY. 2403 01:49:45,313 --> 01:49:48,249 HE'S AN EXPERT IN DATA SCIENCE 2404 01:49:48,249 --> 01:49:50,452 AND BIOINFORMATICS AND 2405 01:49:50,452 --> 01:49:51,419 STRUCTURED BIOLOGY AND 2406 01:49:51,419 --> 01:49:53,488 STRUCTURED DRUG DISCOVERY FOR 2407 01:49:53,488 --> 01:49:56,825 ONCOLOGY AND DIRECTOR OF THE 2408 01:49:56,825 --> 01:49:58,727 DATABASE BANK AND ALSO WITHIN 2409 01:49:58,727 --> 01:50:02,330 RUTGERS THE STATE UNIVERSITY OF 2410 01:50:02,330 --> 01:50:04,165 NEW JERSEY SERVES AS UNIVERSITY 2411 01:50:04,165 --> 01:50:10,772 PROFESSOR AND HENRY RUTGER CHAIR 2412 01:50:10,772 --> 01:50:14,609 AND THE CANCER RESEARCH PROGRAM 2413 01:50:14,609 --> 01:50:15,977 CO-LEADER WITHIN THE INSTITUTE 2414 01:50:15,977 --> 01:50:16,878 OF NEW JERSEY. 2415 01:50:16,878 --> 01:50:19,180 HE HAS PUBLISHED EXTENSIVELY IN 2416 01:50:19,180 --> 01:50:23,151 DATA SCIENCE AND BIOINFORMATICS 2417 01:50:23,151 --> 01:50:24,152 AND ARTIFICIAL INTELLIGENCE, 2418 01:50:24,152 --> 01:50:26,721 MACHINE LEARNING AND CLINICAL 2419 01:50:26,721 --> 01:50:27,255 ONCOL 2420 01:50:27,255 --> 01:50:28,056 ONCOLOGY. 2421 01:50:28,056 --> 01:50:30,058 HE'LL TALK ABOUT PROTEIN DATA 2422 01:50:30,058 --> 01:50:35,063 BANK FROM THE GLOBAL PANDEMIC TO 2423 01:50:35,063 --> 01:50:39,167 mRNA VACCINES AN PAXLOVID. 2424 01:50:39,167 --> 01:50:49,511 THE FLOOR IS YOURS. 2425 01:51:20,742 --> 01:51:22,744 >> YOU SHOULD NOW SEE MY SLIDE. 2426 01:51:22,744 --> 01:51:25,113 VERY GRATEFUL FOR YOUR KIND 2427 01:51:25,113 --> 01:51:31,986 INTRODUCTION AND THE HONOR OF 2428 01:51:31,986 --> 01:51:38,293 PRESENTING IN THE HONOR OF 2429 01:51:38,293 --> 01:51:39,093 PRESENTING IN THIS IN SILICO 2430 01:51:39,093 --> 01:51:41,062 DRUG DISCOVERY WORKSHOP. 2431 01:51:41,062 --> 01:51:42,463 TODAY I'LL PRESENT REAL WORLD 2432 01:51:42,463 --> 01:51:43,631 EXAMPLES OF THE IMPORTANCE OF 2433 01:51:43,631 --> 01:51:45,066 DATA PRESERVATION AND FAIR 2434 01:51:45,066 --> 01:51:47,202 SHARING. 2435 01:51:47,202 --> 01:51:50,238 IN THIS TALK ENTITLED PROTEIN 2436 01:51:50,238 --> 01:51:54,509 DATA BANK FROM TWO EPIDEMICS TO 2437 01:51:54,509 --> 01:51:56,811 THE PANDEMIC AND mRNA VACCINES 2438 01:51:56,811 --> 01:51:58,012 AND PAXLOVID. 2439 01:51:58,012 --> 01:52:00,949 MY PLAN IS TO TELL YOU ABOUT THE 2440 01:52:00,949 --> 01:52:02,817 CRITICAL ROLE THAT STRUCTURAL 2441 01:52:02,817 --> 01:52:05,486 BIOLOGISTS AND THE PDB TOGETHER 2442 01:52:05,486 --> 01:52:09,057 PLAYED IN FIGHTING THE COVID-19 2443 01:52:09,057 --> 01:52:13,494 PANDEMIC. 2444 01:52:13,494 --> 01:52:16,731 THE PUNCH LINE WAS PUT FORWARD 2445 01:52:16,731 --> 01:52:19,500 FROM DR. FAUCI HE CITED THE 2446 01:52:19,500 --> 01:52:21,035 IMPORTANCE OF 3-D VIRUS 2447 01:52:21,035 --> 01:52:22,670 STRUCTURE INFORMING AND SAID 2448 01:52:22,670 --> 01:52:29,477 SHOW ME A PERSON WHO WAS 2449 01:52:29,477 --> 01:52:33,214 VACCINATED, GOT COVID AND TOOK 2450 01:52:33,214 --> 01:52:38,853 PAX PAXLOVID AND DIED AND I'LL 2451 01:52:38,853 --> 01:52:40,722 COVER THE DATA BANK PARTNERSHIP 2452 01:52:40,722 --> 01:52:44,158 AND THE RCSP DATA BANK AND MY 2453 01:52:44,158 --> 01:52:45,693 ORGANIZATION THE U.S. DATA 2454 01:52:45,693 --> 01:52:48,830 CENTER FOR THE WWPDB. 2455 01:52:48,830 --> 01:52:51,666 OUR ONE-STOP SHOP FOR VIRUS 2456 01:52:51,666 --> 01:52:52,634 STRUCTURE DATA AND THEN SPEND 2457 01:52:52,634 --> 01:52:55,536 THE BULK OF MY TIME TALKING 2458 01:52:55,536 --> 01:52:59,274 ABOUT HOW STRUCTURAL BIOLOGISTS 2459 01:52:59,274 --> 01:53:04,545 AND THE PDB PLAY ROLES IN TAMING 2460 01:53:04,545 --> 01:53:05,480 OF THE COVID-19 PANDEMIC. 2461 01:53:05,480 --> 01:53:10,518 AND CLOSE WITH A SOBERING 2462 01:53:10,518 --> 01:53:20,695 PRESCRIPT. 2463 01:53:21,863 --> 01:53:25,166 THE PDB HAS BEEN CONTINUOUSLY 2464 01:53:25,166 --> 01:53:28,670 FUNDED AND FOUR OF THE EARLY 2465 01:53:28,670 --> 01:53:31,506 STRUCTURES ARE ILLUSTRATED IN 2466 01:53:31,506 --> 01:53:33,074 THE RIGHT PANEL. 2467 01:53:33,074 --> 01:53:37,145 IN BIOLOGY, FUNCTION FOLLOWS. 2468 01:53:37,145 --> 01:53:41,115 THIS VARIATION ON THE FAMOUS 2469 01:53:41,115 --> 01:53:43,484 PHRASE EXPLAINS SUCCINCTLY THE 2470 01:53:43,484 --> 01:53:46,988 FUNCTION OF A MACRO MOLECULE IS 2471 01:53:46,988 --> 01:53:49,958 DETERMINED BY ITS SHAPE AND 3-D 2472 01:53:49,958 --> 01:53:50,525 STRUCTURE. 2473 01:53:50,525 --> 01:53:53,661 SINCE ITS INCEPTION THE ARCHIVE 2474 01:53:53,661 --> 01:53:58,833 HAS BECOME THE SINGLE GLOBAL 2475 01:53:58,833 --> 01:54:00,768 RESOURCE FOR ATOMIC LEVEL 3-D 2476 01:54:00,768 --> 01:54:01,936 VIRUS STRUCTURE INFORMATION. 2477 01:54:01,936 --> 01:54:04,605 AT PRESENT THE PDB HOUSES MORE 2478 01:54:04,605 --> 01:54:08,042 THAN 225,000 STRUCTURES OF 2479 01:54:08,042 --> 01:54:09,510 PROTEINS, NUCLEIC ACIDS AND 2480 01:54:09,510 --> 01:54:12,113 VIRUSES AND MACRO MOLECULAR 2481 01:54:12,113 --> 01:54:12,613 MACHINES. 2482 01:54:12,613 --> 01:54:13,514 RECOGNIZING THE GLOBAL 2483 01:54:13,514 --> 01:54:16,617 IMPORTANCE OF THE PDB IT'S BEEN 2484 01:54:16,617 --> 01:54:19,887 MANAGED JOINTLY SINCE 2003 BY 2485 01:54:19,887 --> 01:54:21,255 THE WORLDWIDE PROTEIN DATABASE 2486 01:54:21,255 --> 01:54:23,358 PARTNERSHIP AND DATA CENTERS IN 2487 01:54:23,358 --> 01:54:25,493 THE U.S., EUROPE, JAPAN AND THE 2488 01:54:25,493 --> 01:54:27,061 PEOPLE'S REPUBLIC OF CHINA AND 2489 01:54:27,061 --> 01:54:35,069 TWO REPOSITORIES FOR MICROSCOPY 2490 01:54:35,069 --> 01:54:36,938 AND THE DATA BANK SERVES AS A 2491 01:54:36,938 --> 01:54:39,073 DATA CENTER FOR THE PARTNERSHIP 2492 01:54:39,073 --> 01:54:45,079 AND AS THE DESIGNATED WW PDB 2493 01:54:45,079 --> 01:54:47,849 KEEPER FOR THE DATA BANK AND 2494 01:54:47,849 --> 01:54:50,084 HEADQUARTERED AT RUTGERS IN NEW 2495 01:54:50,084 --> 01:54:53,121 JERSEY WITH THREE ADDITIONAL 2496 01:54:53,121 --> 01:54:54,389 PERFORMANCE SITES AT U.C. SAN 2497 01:54:54,389 --> 01:54:56,290 DIEGO AND SAN FRANCISCO. 2498 01:54:56,290 --> 01:54:58,226 WE OPERATE FIVE INTERLOCKING 2499 01:54:58,226 --> 01:54:59,527 SERVICES THAT CONVERT GLOBAL 2500 01:54:59,527 --> 01:55:02,830 DATA INTO GLOBAL KNOWLEDGE FOR 2501 01:55:02,830 --> 01:55:03,698 THE PUBLIC GOOD. 2502 01:55:03,698 --> 01:55:05,700 ONCE 3-D VIRUS STRUCTURE DATA 2503 01:55:05,700 --> 01:55:09,370 ARE ARCHIVED IN THE PDB, OUR 2504 01:55:09,370 --> 01:55:14,475 RESEARCH FOCUS WEB PORTAL SERVES 2505 01:55:14,475 --> 01:55:18,679 AS A RESOURCE FOR EXPERIMENTALLY 2506 01:55:18,679 --> 01:55:20,181 DETERMINED PDB STRUCTURES 2507 01:55:20,181 --> 01:55:27,422 ALONGSIDE 1 MILLION MODELS. 2508 01:55:27,422 --> 01:55:31,225 AND PUBLIC DOMAIN INFORMATION TO 2509 01:55:31,225 --> 01:55:32,960 MANY RESEARCHERS AND THEIR 2510 01:55:32,960 --> 01:55:34,562 TRAINEES AND EDUCATORS AND THEIR 2511 01:55:34,562 --> 01:55:36,130 STUDENTS AROUND THE WORLD AT NO 2512 01:55:36,130 --> 01:55:36,898 COST WITH NO LIMITATIONS ON 2513 01:55:36,898 --> 01:55:39,067 USAGE. 2514 01:55:39,067 --> 01:55:42,336 FOR THE AVOIDANCE THE PROMINENCE 2515 01:55:42,336 --> 01:55:44,172 AND RELIABILITY OF BOTH TYPES OF 2516 01:55:44,172 --> 01:55:47,208 3-D VIRUS STRUCTURE DATA 2517 01:55:47,208 --> 01:55:49,510 EXPERIMENTALLY DETERMINE AND 2518 01:55:49,510 --> 01:55:51,212 COMPUTATIONALLY PREDICT THE 2519 01:55:51,212 --> 01:55:53,414 CLEARLY IDENTIFIED ON THE 2520 01:55:53,414 --> 01:55:53,648 WEBSITE. 2521 01:55:53,648 --> 01:55:54,615 PDB DATA ARE ESSENTIAL FOR 2522 01:55:54,615 --> 01:55:57,418 RESPONDING TO EMERGING VIRUSES. 2523 01:55:57,418 --> 01:56:02,123 SINCE THE EARLY 2000s THE WORLD 2524 01:56:02,123 --> 01:56:05,226 HAS FACED DOWN THREE MAJOR 2525 01:56:05,226 --> 01:56:08,262 OUTBREAKS OF CORONAVIRUS 2526 01:56:08,262 --> 01:56:13,367 OUTBREAKS THAT JUMP THE SPECIES 2527 01:56:13,367 --> 01:56:15,937 TO HUMAN AND THEY KILLED JUST 2528 01:56:15,937 --> 01:56:21,075 OVER 800 PEOPLE BETWEEN 2002 AND 2529 01:56:21,075 --> 01:56:21,342 2004. 2530 01:56:21,342 --> 01:56:28,416 HERE AFTER I'LL REFER TO SARS 2531 01:56:28,416 --> 01:56:34,322 COV TO SARS COV1 TO AVOID 2532 01:56:34,322 --> 01:56:36,457 CONFUSION AND TO DATE IT'S 2533 01:56:36,457 --> 01:56:38,259 KILLED MORE THAN 900 INDIVIDUALS 2534 01:56:38,259 --> 01:56:43,431 AND REMAINS A PUBLIC HEALTH 2535 01:56:43,431 --> 01:56:45,066 THREAT IN THE MIDDLE EAST AND 2536 01:56:45,066 --> 01:56:46,267 ASIA. 2537 01:56:46,267 --> 01:56:47,468 DESPITE THE TWO WARNING SIGNS 2538 01:56:47,468 --> 01:56:49,871 FEW COUNTRIES PREPARED FOR THE 2539 01:56:49,871 --> 01:56:53,040 POSSIBILITY OF A MUCH MORE 2540 01:56:53,040 --> 01:56:55,109 CORONAVIRUS EPIDEMIC. 2541 01:56:55,109 --> 01:56:57,078 FORTUNATELY FOR THE GLOBAL 2542 01:56:57,078 --> 01:56:59,180 COMMUNITY, STRUCTURAL BIOLOGISTS 2543 01:56:59,180 --> 01:57:00,414 AND THE PDB LAID THE GROUND WORK 2544 01:57:00,414 --> 01:57:04,318 FOR A SUCCESSFUL RESPONSE TO THE 2545 01:57:04,318 --> 01:57:05,486 INEVITABLE THIRD CORONAVIRUS 2546 01:57:05,486 --> 01:57:06,521 WAVE. 2547 01:57:06,521 --> 01:57:08,923 THE EFFECTIVE VACCINES WERE 2548 01:57:08,923 --> 01:57:12,493 DESIGN AND ANTIVIRAL AGENTS WERE 2549 01:57:12,493 --> 01:57:13,728 DEVELOPED WITH THE BENEFIT OF 2550 01:57:13,728 --> 01:57:17,899 OPEN ACCESS TO PDB STRUCTURES OF 2551 01:57:17,899 --> 01:57:20,902 CARS COV2 AND MERS AND PROTEINS. 2552 01:57:20,902 --> 01:57:23,271 SARS COV2 THE UNSEEN ENEMY IS 2553 01:57:23,271 --> 01:57:25,740 SHOWN ON THE RIGHT WITHIN A 2554 01:57:25,740 --> 01:57:29,043 RESPIRATORY DROP LET IN A 2555 01:57:29,043 --> 01:57:31,913 HAUNTING PAINTING BY THE AUTHOR 2556 01:57:31,913 --> 01:57:35,816 OF THE RCSV PDB WHO AUTHORED THE 2557 01:57:35,816 --> 01:57:37,485 SERIES MORE THAN 20 YEARS. 2558 01:57:37,485 --> 01:57:40,521 THE ORGANIZATION OF THE GENOME 2559 01:57:40,521 --> 01:57:43,057 IS DEPICTED SCHEMATICALLY. 2560 01:57:43,057 --> 01:57:44,825 ALL CORONAVIRUS GENOMES ARE VERY 2561 01:57:44,825 --> 01:57:49,830 LONG, SINGLE STRANDED POSITIVE 2562 01:57:49,830 --> 01:58:00,241 SENSE POLY A DEDEN DEDENNY -- 2563 01:58:00,241 --> 01:58:02,610 PROTEINS ARE EXPRESSED WITHIN A 2564 01:58:02,610 --> 01:58:03,578 PAIR OF PROTEINS AND THERE'S 2565 01:58:03,578 --> 01:58:06,814 EXCISED FROM THE POLY PROTEIN BY 2566 01:58:06,814 --> 01:58:12,587 TWO SARS COV2 PROTEASES NON 2567 01:58:12,587 --> 01:58:15,823 STRUCTURE PROTEINS THEMSELVES. 2568 01:58:15,823 --> 01:58:19,961 THE PROTINNASE HAS ARROWS. 2569 01:58:19,961 --> 01:58:22,930 AND MORE IMPORTANT MAIN PROTEASE 2570 01:58:22,930 --> 01:58:25,499 CUTS AT 10 SIGHT WITH LIGHT BLUE 2571 01:58:25,499 --> 01:58:27,034 INVERTED TRIANGLES. 2572 01:58:27,034 --> 01:58:28,736 THOUGH WE DIDN'T KNOW IT AT THE 2573 01:58:28,736 --> 01:58:32,306 TIME, PATHOGEN PREPAREDNESS 2574 01:58:32,306 --> 01:58:38,145 BEGAN IN THE 2000s WITH WORK ON 2575 01:58:38,145 --> 01:58:41,849 SARS COVE 1 AND THE MAIN 2576 01:58:41,849 --> 01:58:45,553 PROTEASE OR EMPRO. 2577 01:58:45,553 --> 01:58:48,956 AND THEY USED THE ALPHABET. 2578 01:58:48,956 --> 01:58:55,129 WITHOUT THE MAIN PROTEASE IT 2579 01:58:55,129 --> 01:58:56,597 CANNOT CLEAVE STOPPING AN 2580 01:58:56,597 --> 01:58:57,331 INFECTION IN ITS TRACKS. 2581 01:58:57,331 --> 01:59:04,572 THIS STRUCTURE WAS DETERMINED IN 2582 01:59:04,572 --> 01:59:09,076 2003 WE MY COMPANY FCX 2583 01:59:09,076 --> 01:59:09,443 PHARMACEUTICALS. 2584 01:59:09,443 --> 01:59:13,281 IT'S THE ONLY PDB STRUCTURE WITH 2585 01:59:13,281 --> 01:59:15,583 THE PRIMARY PUBLICATION 2586 01:59:15,583 --> 01:59:19,253 APPEARING IN THE NEW YORK TIMES. 2587 01:59:19,253 --> 01:59:20,221 MY COMPANY RELEASED THE 2588 01:59:20,221 --> 01:59:22,356 STRUCTURE TO THE PUBLIC SO WORK 2589 01:59:22,356 --> 01:59:25,459 ON COUNTER MEASURES COULD BEGIN 2590 01:59:25,459 --> 01:59:25,860 WITHOUT DELAY. 2591 01:59:25,860 --> 01:59:30,031 TODAY 98 CRYSTAL STRUCTURES OF 2592 01:59:30,031 --> 01:59:30,898 THE PROTEASE HAVE BEEN 2593 01:59:30,898 --> 01:59:32,900 CONTRIBUTED TO THE ARCHIVE. 2594 01:59:32,900 --> 01:59:36,270 THE FIRST PDB STRUCTURE OF THE 2595 01:59:36,270 --> 01:59:39,340 ENTIRE EXTRA CELLULAR STRUCTURE 2596 01:59:39,340 --> 01:59:43,110 OF THE TRIMERIC STRIKE PROTEIN 2597 01:59:43,110 --> 01:59:45,680 WAS DETERMINED USING SINGLE 2598 01:59:45,680 --> 01:59:51,052 PARTICLE MICROSCOPY. 2599 01:59:51,052 --> 01:59:53,854 TO DATE MORE THAN 70 SARS COV1 2600 01:59:53,854 --> 01:59:59,627 STRUCTURES HAVE SBN CONTRBEEN CD 2601 01:59:59,627 --> 02:00:01,495 AND CONFIRMATIONS AND COMPLEXES 2602 02:00:01,495 --> 02:00:05,333 WITH THE CELLULAR RECEPTOR AND 2603 02:00:05,333 --> 02:00:06,834 CONVERTING ENZYME 2 AND 2604 02:00:06,834 --> 02:00:08,736 VARIATION ANTIBODY FAB 2605 02:00:08,736 --> 02:00:09,070 FRAGMENTS. 2606 02:00:09,070 --> 02:00:13,274 THE FIRST PDB STRUCTURE OF MERS 2607 02:00:13,274 --> 02:00:15,676 PROTEIN WAS THAT OF THE ACE 2 2608 02:00:15,676 --> 02:00:17,111 BINDING DOMAIN OF THE SPIKE 2609 02:00:17,111 --> 02:00:18,946 PROTEIN SHOWN HERE IN THE RIGHT 2610 02:00:18,946 --> 02:00:21,082 PANEL IN A RIBBON 2611 02:00:21,082 --> 02:00:22,416 REPRESENTATION. 2612 02:00:22,416 --> 02:00:24,952 TO DATE, FOR MERS MORE THAN 70 2613 02:00:24,952 --> 02:00:27,054 SPIKE PROTEIN RELATED STRUCTURES 2614 02:00:27,054 --> 02:00:29,490 HAVE BEEN CONTRIBUTED TO THE 2615 02:00:29,490 --> 02:00:33,494 ARCHIVE ALONGSIDE MORE THAN 40M 2616 02:00:33,494 --> 02:00:36,564 PRO STRUCTURES. 2617 02:00:36,564 --> 02:00:44,839 I'LL RETURN TO THE SARS AND MERS 2618 02:00:44,839 --> 02:00:45,473 PROTEIN STRUCTURES LATER BECAUSE 2619 02:00:45,473 --> 02:00:48,476 THEY PLAYED CRITICAL ROLES IN 2620 02:00:48,476 --> 02:00:49,910 THE VACCINE DESIGN. 2621 02:00:49,910 --> 02:00:53,481 SINCE THEY WERE DEPOSITED IN THE 2622 02:00:53,481 --> 02:00:58,519 PDB THE RCSP PROTEIN DATA BANK 2623 02:00:58,519 --> 02:00:59,854 HAS BEEN RESPONDING TO THE 2624 02:00:59,854 --> 02:01:00,121 PANDEMIC. 2625 02:01:00,121 --> 02:01:03,057 WITH OUR PARTNERS WE PRIORITIZED 2626 02:01:03,057 --> 02:01:05,826 COMPLETE DEPOSITION, RIGOROUS 2627 02:01:05,826 --> 02:01:07,528 VALIDATION AND EXPERT BIO 2628 02:01:07,528 --> 02:01:08,929 CURATION OF COVID-19 RELATED 2629 02:01:08,929 --> 02:01:12,333 PROTEIN STRUCTURES THAT STRONGLY 2630 02:01:12,333 --> 02:01:13,033 ENCOURAGE IMMEDIATE 2631 02:01:13,033 --> 02:01:13,667 PREPUBLICATION RELEASE OF THESE 2632 02:01:13,667 --> 02:01:23,778 DATA. 2633 02:01:29,950 --> 02:01:31,452 WE CREATED A CORONAVIRUS PAGE ON 2634 02:01:31,452 --> 02:01:34,221 OUR WEB PORTAL UPDATED WEEKLY 2635 02:01:34,221 --> 02:01:36,090 WITH EACH RELEASE CYCLE ADDING 2636 02:01:36,090 --> 02:01:37,992 APPROXIMATELY 300 NEW STRUCTURES 2637 02:01:37,992 --> 02:01:41,495 OF ANY STRIFE TO THE ARCHIVE 2638 02:01:41,495 --> 02:01:44,498 EVERY WEDNESDAY AND DEVELOPED 2639 02:01:44,498 --> 02:01:46,867 OUTREACH RESOURCES SUCH AS THE 2640 02:01:46,867 --> 02:01:49,503 ANIMATED HAND WASHING VIDEO 2641 02:01:49,503 --> 02:01:51,372 VIEWED MORE THAN 500,000 TIMES 2642 02:01:51,372 --> 02:01:55,376 AND PDB MOLECULE OF THE MONTH 2643 02:01:55,376 --> 02:01:57,478 ARTICLES AUTHORED BY DAVID 2644 02:01:57,478 --> 02:02:01,115 GOODSELL FEATURING THE PROTEASES 2645 02:02:01,115 --> 02:02:06,086 AND THE POLYMERASE AND SPIKE 2646 02:02:06,086 --> 02:02:06,854 PROTEINS. 2647 02:02:06,854 --> 02:02:09,023 THE RIGHT PANEL SUMMARIZES OUR 2648 02:02:09,023 --> 02:02:11,158 NEARLY COMPLETE ATOM BEING LEVEL 2649 02:02:11,158 --> 02:02:12,993 3-D KNOWLEDGE OF SARS COV2 2650 02:02:12,993 --> 02:02:16,397 PROTEIN STRUCTURES EMBODIED IN 2651 02:02:16,397 --> 02:02:20,468 MORE THAN 4500 PDB STRUCTURES 2652 02:02:20,468 --> 02:02:20,968 RELEASED. 2653 02:02:20,968 --> 02:02:22,536 THAT REPRESENTS NEARLY 2% OF THE 2654 02:02:22,536 --> 02:02:25,206 ENTIRE PDB ARCHIVE. 2655 02:02:25,206 --> 02:02:27,508 THE LEFT PANEL SHOWS AN ARTIST 2656 02:02:27,508 --> 02:02:31,445 RENDITION OF THE LIPID BILAYER 2657 02:02:31,445 --> 02:02:33,514 OF SARS COV2 FUSING WITH THE 2658 02:02:33,514 --> 02:02:35,850 PLASMA MEMBRANE OF A CELL IN 2659 02:02:35,850 --> 02:02:36,684 ANOTHER HAUNTING PAINTING FROM 2660 02:02:36,684 --> 02:02:39,086 DAVID GOODSELL. 2661 02:02:39,086 --> 02:02:41,522 THE VIRAL ENTRY PROCESS IS 2662 02:02:41,522 --> 02:02:44,758 BLOCKED BY THE IMMUNE RESPONSE 2663 02:02:44,758 --> 02:02:46,861 TO mRNA VACCINES THAT COME ON 2664 02:02:46,861 --> 02:02:47,628 TOO SOON. 2665 02:02:47,628 --> 02:02:50,498 THE VERY FIRST PDB STRUCTURE OF 2666 02:02:50,498 --> 02:02:56,837 A SARS COV2 SPIKE PROTEIN ID6VSB 2667 02:02:56,837 --> 02:02:58,305 WAS RELEASED IN 2020. 2668 02:02:58,305 --> 02:03:01,509 THERE'S ABOUT 1900 SPIKE PROTEIN 2669 02:03:01,509 --> 02:03:05,679 RELATED STRUCK STURZ IN THE PDB. 2670 02:03:05,679 --> 02:03:07,982 LIKE EARLIER STRUCTURES THEY 2671 02:03:07,982 --> 02:03:09,517 PROVIDE IMPORTANT INSIGHTS INTO 2672 02:03:09,517 --> 02:03:12,486 RECEPTOR BINDING, FUSION WITH 2673 02:03:12,486 --> 02:03:18,225 THE BILAYER AND THE PLASMA MANY 2674 02:03:18,225 --> 02:03:20,160 MEMBRANE AND ANTIBODY DESIGN. 2675 02:03:20,160 --> 02:03:21,829 THE FIRST mRNA VACCINE DESIGNS 2676 02:03:21,829 --> 02:03:25,366 WERE GENERATED IN JANUARY 2020 2677 02:03:25,366 --> 02:03:29,837 BASED ON STRUCTURES OF THE COV1 2678 02:03:29,837 --> 02:03:33,073 PROTEINS SIMILAR IN AMINO ACID 2679 02:03:33,073 --> 02:03:35,910 AND 3-D STRUCTURE TO THEIR SARS 2680 02:03:35,910 --> 02:03:37,511 COV2 COUNTERPART. 2681 02:03:37,511 --> 02:03:42,983 WHEN DESIGNING MODERNA'S 1273 2682 02:03:42,983 --> 02:03:45,119 VACCINE ENCODING THE SPIKE 2683 02:03:45,119 --> 02:03:47,755 PROTEIN, RESEARCHERS AND THE FOR 2684 02:03:47,755 --> 02:03:49,523 PROFIT COMPANY REPORTED RELIANCE 2685 02:03:49,523 --> 02:03:52,893 ON PDB STRUCTURES OF DOUBLE 2686 02:03:52,893 --> 02:03:57,031 PROLEAN FORMS OF THE SARS COV1 2687 02:03:57,031 --> 02:03:59,133 AND MERS SPIKE PROTEINS. 2688 02:03:59,133 --> 02:04:01,535 THE DOUBLE PROLENE MUTATION 2689 02:04:01,535 --> 02:04:03,237 STABILIZES BOTH PROTEINS IN A 2690 02:04:03,237 --> 02:04:05,673 HIGHLY IMMUNOGENIC PREFUSION 2691 02:04:05,673 --> 02:04:07,274 CONFIRMATION. 2692 02:04:07,274 --> 02:04:09,543 I BELIEVE THE VACCINE DESIGNERS 2693 02:04:09,543 --> 02:04:13,447 AT BIO EN TECH USED THE SAME 2694 02:04:13,447 --> 02:04:15,082 BECAUSE THEY'RE INAUGURAL 2695 02:04:15,082 --> 02:04:18,686 VACCINE ALSO ENCODED A SARS COV2 2696 02:04:18,686 --> 02:04:21,488 PROTEIN BEARING THE IDENTICAL 2697 02:04:21,488 --> 02:04:22,056 PROLENE CHANGE. 2698 02:04:22,056 --> 02:04:23,190 THE REST OF THE STORY IS WELL 2699 02:04:23,190 --> 02:04:25,259 KNOWN TO YOU. 2700 02:04:25,259 --> 02:04:27,461 TWO ESSENTIALLY IDENTICAL mRNA 2701 02:04:27,461 --> 02:04:28,495 VACCINES AGAINST COVID-19 WERE 2702 02:04:28,495 --> 02:04:29,530 BROUGHT TO THE MARKET IN RECORD 2703 02:04:29,530 --> 02:04:31,498 TIME. 2704 02:04:31,498 --> 02:04:33,167 TO DATE, THESE VACCINES AND 2705 02:04:33,167 --> 02:04:36,937 THEIR SUCCESSORS HAVE BEEN 2706 02:04:36,937 --> 02:04:39,540 ADMINISTERS TO MORE THAN 5.5 2707 02:04:39,540 --> 02:04:41,508 MILLION INDIVIDUALS WITH 2708 02:04:41,508 --> 02:04:42,710 IMPRESSIVE RESULTS. 2709 02:04:42,710 --> 02:04:44,778 THE VACCINE DESIGN WITH THE 2710 02:04:44,778 --> 02:04:49,049 BENEFIT OF OPEN ACCESS TO PDB 2711 02:04:49,049 --> 02:04:53,621 STUK TOU STRUCTURES CREDITED WI 2712 02:04:53,621 --> 02:04:57,524 SAVING MILLIONS OF LIVES AND YOU 2713 02:04:57,524 --> 02:05:00,828 HAVE TO LOVE THE PROTEIN DATA 2714 02:05:00,828 --> 02:05:01,528 BANK. 2715 02:05:01,528 --> 02:05:03,230 I'LL NOW TURN FROM THE SPIKE 2716 02:05:03,230 --> 02:05:07,368 PROTEIN AND VACCINE DESIGN TO 2717 02:05:07,368 --> 02:05:09,937 DISCOVERY OF SMALL MOLECULE 2718 02:05:09,937 --> 02:05:13,273 ANTISARS DRUGS. 2719 02:05:13,273 --> 02:05:16,944 DRUG PRE PURPOSING PROVED TO BE 2720 02:05:16,944 --> 02:05:21,081 IMPORTANT FOR THE POLYMERASE 2721 02:05:21,081 --> 02:05:24,585 SHOWN ID6YYT. 2722 02:05:24,585 --> 02:05:27,988 THE HOLLOW ENZYME CONSISTS OF 2723 02:05:27,988 --> 02:05:31,125 THE SMALLER NSP7 AND LARGER 12 2724 02:05:31,125 --> 02:05:35,229 AND LARGER POLYMERASE SUB UNIT 2725 02:05:35,229 --> 02:05:37,531 AND THE MOLECULAR CHOP STICKS 2726 02:05:37,531 --> 02:05:44,104 HOLDING THE DUPLEX. 2727 02:05:44,104 --> 02:05:46,507 GILEAD DEVELOPED A TREATMENT FOR 2728 02:05:46,507 --> 02:05:48,475 HEPATITIS C AND RECEIVED 2729 02:05:48,475 --> 02:05:53,514 EMERGENCY USE AUTHORIZATION FOR 2730 02:05:53,514 --> 02:05:57,951 SARS COV2 FROM US FDA AND 2731 02:05:57,951 --> 02:06:01,722 ANOTHER WAS INCORPORATED IN THE 2732 02:06:01,722 --> 02:06:04,958 PRODUCT AND THIS EXPLAINS THE 2733 02:06:04,958 --> 02:06:07,127 MECHANISM OF THE ACTION AT THE 2734 02:06:07,127 --> 02:06:13,500 ATOMIC LEVEL OF THE REPURPOSED 2735 02:06:13,500 --> 02:06:17,438 DRUG IN 3-D AND THIS ONE 2736 02:06:17,438 --> 02:06:19,473 DEVELOPED FOR THE EQUINE 2737 02:06:19,473 --> 02:06:21,108 ENCEPHALITIS VIRUS NOW A PUBLIC 2738 02:06:21,108 --> 02:06:23,410 HEALTH THREAT IN THE UNITED 2739 02:06:23,410 --> 02:06:24,278 STATES. 2740 02:06:24,278 --> 02:06:25,546 IT RECEIVED EIGHT EMERGENCY USE 2741 02:06:25,546 --> 02:06:27,715 AUTHORIZATION FOR SARS COV2 FROM 2742 02:06:27,715 --> 02:06:33,120 THE U.S. FDA IN 2021. 2743 02:06:33,120 --> 02:06:37,091 THE RIGHT PANEL SHOWS PDB OZU 2744 02:06:37,091 --> 02:06:38,358 INCORPORATED INTO THIS RNA 2745 02:06:38,358 --> 02:06:40,694 PRODUCT OF THE HOLLOW ENZYME. 2746 02:06:40,694 --> 02:06:42,796 LIKE REMDESIVIR THIS SHOWS THE 2747 02:06:42,796 --> 02:06:44,031 MECHANISM OF ACTION OF THE 2748 02:06:44,031 --> 02:06:46,233 REPURPOSED DRUG AT THE ATOMIC 2749 02:06:46,233 --> 02:06:48,268 LEVEL IN 3-D. 2750 02:06:48,268 --> 02:06:50,304 MOVING ON I'LL NOW DESCRIBE 2751 02:06:50,304 --> 02:06:51,972 STRUCTURE GUIDED DRUG DISCOVERY 2752 02:06:51,972 --> 02:06:56,577 FOR SARS COV2 MAIN PROTEASE THE 2753 02:06:56,577 --> 02:06:58,145 ACHILLES HEEL OF THE VIRUS. 2754 02:06:58,145 --> 02:07:08,388 IT'S THE TARGET OF PFIZER'S DRUG 2755 02:07:08,388 --> 02:07:10,824 PAXLOVID A FIXED DOSE. 2756 02:07:10,824 --> 02:07:14,762 IT HOUSES MORE THAN 1500 CRYSTAL 2757 02:07:14,762 --> 02:07:16,997 STRUCTURES OF SARS COV2 EMPRO. 2758 02:07:16,997 --> 02:07:20,000 THE FIRST OF ID UNIFORM 7 2759 02:07:20,000 --> 02:07:23,771 CONTRIBUTED IN JANUARY 2020 BY 2760 02:07:23,771 --> 02:07:26,607 COLLEAGUES BASED IN SHANGHAI 2761 02:07:26,607 --> 02:07:28,475 APPROXIMATELY TWO WEEKS AFTER 2762 02:07:28,475 --> 02:07:31,612 THE GENOME BECAME PUBLICALLY 2763 02:07:31,612 --> 02:07:33,647 AVAILABLE THE ORIGIN STORY 2764 02:07:33,647 --> 02:07:36,617 BEGINS IN THE EARLY 2000s WITH 2765 02:07:36,617 --> 02:07:43,090 DEVELOPMENT OF PF-008321 SHOWN 2766 02:07:43,090 --> 02:07:47,227 IN THE RIGHT PANEL INHIBITING 2767 02:07:47,227 --> 02:07:48,028 SARS COV1 DRUG. 2768 02:07:48,028 --> 02:07:52,432 THIS WAS INTENDED AS AN 2769 02:07:52,432 --> 02:07:57,504 INJECTABLE ANTIVIRAL AGENT FOR 2770 02:07:57,504 --> 02:07:59,339 ACUTELY ILL SARS COV1 PATIENTS 2771 02:07:59,339 --> 02:08:09,817 AND THIS INITIATED IN 2003 2772 02:08:12,186 --> 02:08:17,524 FACILITATED BY THIS AND 2773 02:08:17,524 --> 02:08:19,660 REGRETTABLY BUT UNDERSTANDABLY 2774 02:08:19,660 --> 02:08:22,462 PFIZER ALTERED THE PROJECT 2775 02:08:22,462 --> 02:08:26,900 BECAUSE SARS COV1 AND THE 2776 02:08:26,900 --> 02:08:28,468 COMPANY JUDGED THERE'D BE NO 2777 02:08:28,468 --> 02:08:30,170 COMMERCIAL MARKET FOR THE DRUG. 2778 02:08:30,170 --> 02:08:33,140 AS THE PANDEMIC HIT, PFIZER 2779 02:08:33,140 --> 02:08:39,680 QUICKLY REACT VATED THE SARS 2780 02:08:39,680 --> 02:08:41,982 COV1 AGENT FOR OUT PATIENT USE 2781 02:08:41,982 --> 02:08:48,856 FACILITATED BY OPEN ACCESS TO, 2782 02:08:48,856 --> 02:08:53,527 THE FIRST STRUCTURE I DESCRIBED 2783 02:08:53,527 --> 02:08:54,061 EARLIER. 2784 02:08:54,061 --> 02:08:59,266 THE RESULT WAS NIRMATRELVIR. 2785 02:08:59,266 --> 02:09:03,103 YOU HAVE TO LOVE THE PROTEIN 2786 02:09:03,103 --> 02:09:06,807 DATA BANK. 2787 02:09:06,807 --> 02:09:09,042 AND WHAT WAS SHOWN ON THE 2788 02:09:09,042 --> 02:09:12,012 PREVIOUS SIDE COVALENTLY ACT 2789 02:09:12,012 --> 02:09:14,147 BEING INHIBITORS OF THE PROTEASE 2790 02:09:14,147 --> 02:09:19,720 BINDING TO THE CYSTEINE RESIDUE 2791 02:09:19,720 --> 02:09:22,522 AND IT CATALYZES A SINGLE TURN 2792 02:09:22,522 --> 02:09:25,525 OVER REACTION FORM THEY CAN 2793 02:09:25,525 --> 02:09:27,427 ENACTIVE COVALENT ADAPT IN THE 2794 02:09:27,427 --> 02:09:30,964 RIGHT PANEL PREVENTING PROCESS 2795 02:09:30,964 --> 02:09:32,933 OF THE VIRAL POLY PROTEIN AND 2796 02:09:32,933 --> 02:09:34,434 STOPPING THE INFECTION IN ITS 2797 02:09:34,434 --> 02:09:35,302 TRACK. 2798 02:09:35,302 --> 02:09:38,438 IT'S STRUCTURALLY SIMILAR TO THE 2799 02:09:38,438 --> 02:09:42,743 FIRST GENERATION SARS COV1 EMPRO 2800 02:09:42,743 --> 02:09:44,745 INHIBITOR AND TO BE ADMINISTERED 2801 02:09:44,745 --> 02:09:49,516 AS A PILL TO NEWLY INFECTED 2802 02:09:49,516 --> 02:09:53,020 INDIVIDUALS. 2803 02:09:53,020 --> 02:10:01,528 NIRMATRELVIR IS RAPIDLY 2804 02:10:01,528 --> 02:10:02,896 METABOLIZED AND A POTENT 2805 02:10:02,896 --> 02:10:06,533 INHIBITOR PROLONGS THE HALF LIFE 2806 02:10:06,533 --> 02:10:08,869 BY PREVENTING DEGRADATION AS IT 2807 02:10:08,869 --> 02:10:11,338 PASSES THROUGH THE LIVER. 2808 02:10:11,338 --> 02:10:12,005 PAXLOVID, A FIXED DOSE 2809 02:10:12,005 --> 02:10:15,575 COMBINATION OF THE TWO DRUGS 2810 02:10:15,575 --> 02:10:17,511 RECEIVED EMERGENCY USE 2811 02:10:17,511 --> 02:10:19,379 AUTHORIZATION FROM U.S. FDA IN 2812 02:10:19,379 --> 02:10:19,947 2021. 2813 02:10:19,947 --> 02:10:22,382 LESS THAN TWO YEARS AFTER PUBLIC 2814 02:10:22,382 --> 02:10:25,252 RELEASE OF THE PUBLIC GENOME 2815 02:10:25,252 --> 02:10:25,519 SEQUENCE. 2816 02:10:25,519 --> 02:10:26,954 THE SPEED WITH WHICH PFIZER WAS 2817 02:10:26,954 --> 02:10:29,256 ABLE TO MOVE WAS UNPRECEDENTED. 2818 02:10:29,256 --> 02:10:31,792 AT THE END OF MY TALK I'LL COME 2819 02:10:31,792 --> 02:10:33,527 BACK THAT PFIZER REPORTED IT'S 2820 02:10:33,527 --> 02:10:36,029 ALSO ACTIVE AGAINST THE MAIN 2821 02:10:36,029 --> 02:10:40,400 PROTEASE OF BOTH SARS COV1 AND 2822 02:10:40,400 --> 02:10:41,868 MERS COV. 2823 02:10:41,868 --> 02:10:45,339 THIS LEADS ME TO A SOBERING 2824 02:10:45,339 --> 02:10:45,605 PRESCRIPT. 2825 02:10:45,605 --> 02:10:53,547 THE ARCHIVE HOUSES THE P450 2826 02:10:53,547 --> 02:10:57,117 ISOSOME. 2827 02:10:57,117 --> 02:11:01,555 THE RIGHT PANEL SHOWED THE 2828 02:11:01,555 --> 02:11:02,589 INHIBI 2829 02:11:02,589 --> 02:11:02,856 INHIBITOR. 2830 02:11:02,856 --> 02:11:08,395 IT WAS DISCOVERED BY ABBOTT AS A 2831 02:11:08,395 --> 02:11:10,130 PROTEASE IN HINTER AND EXPLAINS 2832 02:11:10,130 --> 02:11:13,500 WHY IT'S ALSO USED FOR 2833 02:11:13,500 --> 02:11:16,303 PROLONGING THE HALF LIVES OF 2834 02:11:16,303 --> 02:11:17,371 OTHER PROTEASE INHIBITORS 2835 02:11:17,371 --> 02:11:19,339 SUSCEPTIBLE TO DEGRADATION. 2836 02:11:19,339 --> 02:11:24,911 THE FACT THAT THIS INHIBITS THE 2837 02:11:24,911 --> 02:11:26,613 MAIN PROTEASES OF SARS COV2 AND 2838 02:11:26,613 --> 02:11:29,516 COV1 IS NOT SURPRISING. 2839 02:11:29,516 --> 02:11:32,486 IN 2021, FOUR UNDERGRADUATES 2840 02:11:32,486 --> 02:11:36,156 WORKING AT RUTGERS WITH ME IN 2841 02:11:36,156 --> 02:11:39,126 OUR GOAL WAS TO UNDERSTAND EMPRO 2842 02:11:39,126 --> 02:11:41,228 AMINO ACID SEQUENCE AND 3-D 2843 02:11:41,228 --> 02:11:43,397 STRUCTURE AND ACTIVE SIDE AND 2844 02:11:43,397 --> 02:11:45,132 THE DIVERSITY ACROSS ALL KNOWN 2845 02:11:45,132 --> 02:11:46,433 CORONAVIRUSES. 2846 02:11:46,433 --> 02:11:48,969 THIS WORK HAS NOT BEEN PUBLISHED 2847 02:11:48,969 --> 02:11:51,772 SO I'LL SUMMARIZE OUR RESULTS IN 2848 02:11:51,772 --> 02:11:52,039 DETAIL. 2849 02:11:52,039 --> 02:11:55,809 ALL CORONAVIRUS MAIN PROTEASES 2850 02:11:55,809 --> 02:12:04,017 FALL INTO FOUR WITHIN THE 2851 02:12:04,017 --> 02:12:04,551 CORONAVIRUS GENUS. 2852 02:12:04,551 --> 02:12:07,888 TAKE A MOMENT TO NOTE HOW CLOSE 2853 02:12:07,888 --> 02:12:11,425 SARS COV1 AND COV2 ARE TO ONE 2854 02:12:11,425 --> 02:12:12,159 ANOTHER. 2855 02:12:12,159 --> 02:12:14,061 THIS IMAGE SHOWN IN THE RIGHT 2856 02:12:14,061 --> 02:12:23,303 PANEL EXPLAINS HOW THE EMPRO 2857 02:12:23,303 --> 02:12:24,671 ACTIVE SITE CAN BE HAVE 2858 02:12:24,671 --> 02:12:26,606 DESIGNATED SITES. 2859 02:12:26,606 --> 02:12:28,408 THE INVARIANT ACTIVE SITE 2860 02:12:28,408 --> 02:12:30,310 CYSTEINE IN THE RED CIRCLE IS 2861 02:12:30,310 --> 02:12:34,481 LOCATED BETWEEN THE S1 PRIME AND 2862 02:12:34,481 --> 02:12:34,648 S1. 2863 02:12:34,648 --> 02:12:37,150 LIKE FULL LENGTH SEQUENCES, ALL 2864 02:12:37,150 --> 02:12:38,552 CORONAVIRUS MAIN PROTEASE ACTIVE 2865 02:12:38,552 --> 02:12:42,622 SITES FALL INTO ONE OF FOUR 2866 02:12:42,622 --> 02:12:46,560 DIFFERENT GENERA BASED ON 2867 02:12:46,560 --> 02:12:48,061 CLUSTERING OF THE RESIDUES THAT 2868 02:12:48,061 --> 02:12:57,137 LINE THE SITE. 2869 02:12:57,137 --> 02:13:01,508 YOU CAN SEE FROM THE DETAIL OF 2870 02:13:01,508 --> 02:13:03,410 THE CLATOGRAM THEY'RE MORE 2871 02:13:03,410 --> 02:13:06,279 SIMILAR TO ONE ANOTHER THAN THE 2872 02:13:06,279 --> 02:13:09,149 FULL LENGTH AMINO ACID 2873 02:13:09,149 --> 02:13:09,516 SEQUENCES. 2874 02:13:09,516 --> 02:13:11,084 CONCENTRATION OF THE RESIDUES 2875 02:13:11,084 --> 02:13:13,086 LINING THE PLAIN PROTEASE ACTIVE 2876 02:13:13,086 --> 02:13:14,754 SITES OF ALL KNOWN CORONAVIRUSES 2877 02:13:14,754 --> 02:13:16,823 IS NOTHING SHORT OF REMARKABLE. 2878 02:13:16,823 --> 02:13:17,924 EACH COLUMN IN THE RIGHT PANEL 2879 02:13:17,924 --> 02:13:20,060 SHOWS THE FREQUENCY OF EACH OF 2880 02:13:20,060 --> 02:13:23,897 THE 20 POSSIBLE AMINO ACIDS 2881 02:13:23,897 --> 02:13:27,000 LINING THE ACTIVE SITE. 2882 02:13:27,000 --> 02:13:29,503 DARK BLUE IDENTIFIES AMINO ACIDS 2883 02:13:29,503 --> 02:13:31,838 THAT DO NOT OCCUR IN THAT 2884 02:13:31,838 --> 02:13:33,173 POSITION AND THERE'S A LOT OF 2885 02:13:33,173 --> 02:13:37,511 DARK BLUE REFLECTING EMPROACTIVE 2886 02:13:37,511 --> 02:13:42,883 SITES ARE CONSERVED ACROSS ALL 2887 02:13:42,883 --> 02:13:43,216 CORONAVIRUSES. 2888 02:13:43,216 --> 02:13:44,484 FOR EXAMPLE, SARS COV1 AND COV2 2889 02:13:44,484 --> 02:13:48,421 ARE MADE OF THE SAME AMINO ACID 2890 02:13:48,421 --> 02:13:51,191 RESIDUES IN THE SAME POSITIONS. 2891 02:13:51,191 --> 02:13:55,629 THERE ARE 12 AMINO ACID RESIDUES 2892 02:13:55,629 --> 02:13:59,099 INVARIANT ACROSS ALL 2893 02:13:59,099 --> 02:14:00,767 CONFERENCES. 2894 02:14:00,767 --> 02:14:02,636 BASED ON COMPUTER MODELLING 2895 02:14:02,636 --> 02:14:03,637 STRUCTURES HOUSED IN THE PDB YOU 2896 02:14:03,637 --> 02:14:07,440 CAN SEE IN THE RIGHT PANEL THE 2897 02:14:07,440 --> 02:14:14,514 POLY PROTEIN SUBSTRATES MODELLED 2898 02:14:14,514 --> 02:14:19,719 CONSISTENT OF SIX AMINO ACID 2899 02:14:19,719 --> 02:14:25,225 RESIDUES P1, P5 AND CATALYZE 2900 02:14:25,225 --> 02:14:26,993 CLEAVAGE OCCURS IN THE PRIME IN 2901 02:14:26,993 --> 02:14:28,094 EACH OF THE SITES DISTRIBUTED 2902 02:14:28,094 --> 02:14:31,565 THROUGH THE LENGTH OF THE POLY 2903 02:14:31,565 --> 02:14:32,599 PROTEIN. 2904 02:14:32,599 --> 02:14:36,036 THE INVARIANT ACTIVE SITE 2905 02:14:36,036 --> 02:14:37,404 CYSTEINE IS VISIBLE WHERE IT 2906 02:14:37,404 --> 02:14:39,272 BREAKS THE BOND BETWEEN P1 AND 2907 02:14:39,272 --> 02:14:43,476 P1 PRIME. 2908 02:14:43,476 --> 02:14:47,747 >> WE HAVE A FEW MINUTES. 2909 02:14:47,747 --> 02:14:51,184 >> I'LL BE DONE ALMOST 2910 02:14:51,184 --> 02:14:52,819 IMMEDIATELY. 2911 02:14:52,819 --> 02:14:54,788 THE CONSEQUENCES ARE ALSO HIGHLY 2912 02:14:54,788 --> 02:14:55,055 CONSERVED. 2913 02:14:55,055 --> 02:14:56,923 THIS SHOWS THE FREQUENCY OF EACH 2914 02:14:56,923 --> 02:14:59,659 OF THE POSSIBLE AMINO ACIDS 2915 02:14:59,659 --> 02:15:01,628 APPEARING P1 THROUGH P1 PRIME. 2916 02:15:01,628 --> 02:15:06,499 TO JUMP TO THE BOTTOM, THERE'S 2917 02:15:06,499 --> 02:15:07,667 DOUBLE EVOLUTIONARY SELECTION 2918 02:15:07,667 --> 02:15:12,973 PRESSURE WORK ON BOTH 3-D 2919 02:15:12,973 --> 02:15:15,442 STRUCTURES OF CONFERENCE, M 2920 02:15:15,442 --> 02:15:16,142 PROACTIVE SIGHT FOUND IN THE 2921 02:15:16,142 --> 02:15:18,945 CORONAVIRUS POLY PROTEIN. 2922 02:15:18,945 --> 02:15:20,313 DURING VIRAL EVOLUTION, ACTIVITY 2923 02:15:20,313 --> 02:15:23,583 OF THE ENZYME HAS TO BE 2924 02:15:23,583 --> 02:15:24,818 PRESERVED AND EACH OF THE TEN 2925 02:15:24,818 --> 02:15:27,187 AMINO ACIDS CANNOT CHANGE 2926 02:15:27,187 --> 02:15:30,523 SIGNIFICANTLY. 2927 02:15:30,523 --> 02:15:32,826 WE THERE PREDICT IT WILL BE 2928 02:15:32,826 --> 02:15:35,395 EFFECTIVE AGAINST MANY OF THE 2929 02:15:35,395 --> 02:15:38,698 KNOWN CORONAVIRUSES NOT JUST 2930 02:15:38,698 --> 02:15:43,403 COV1 AND MERS COV AND BEYOND THE 2931 02:15:43,403 --> 02:15:46,706 WORK WE'RE TESTING THE IN SILICO 2932 02:15:46,706 --> 02:15:48,341 PREDICTIONS AND PROTEASES FROM A 2933 02:15:48,341 --> 02:15:53,513 DIVERSE ARRAY OF CORONAVIRUSES. 2934 02:15:53,513 --> 02:15:55,882 UNDERSTANDING CONSTRAINTS ON THE 2935 02:15:55,882 --> 02:15:57,517 MAIN PROTEASE ACTIVE SITE 2936 02:15:57,517 --> 02:16:01,021 STRUCTURES IN 3-D AND THE CUT 2937 02:16:01,021 --> 02:16:06,426 SITE AMINO ACID TELLS US THE 2938 02:16:06,426 --> 02:16:09,162 FREE MARKET FAILED US. 2939 02:16:09,162 --> 02:16:11,298 IT'S DEFINED AS AN INEFFICIENT 2940 02:16:11,298 --> 02:16:12,399 DISTRIBUTIONS OF GOODS AND 2941 02:16:12,399 --> 02:16:14,734 SERVICES WITHIN THE ECONOMY. 2942 02:16:14,734 --> 02:16:15,302 SUCH PHENOMENA OCCUR WHEN 2943 02:16:15,302 --> 02:16:17,637 INDIVIDUAL INCENTIVES FOR 2944 02:16:17,637 --> 02:16:18,838 RATIONALE ECONOMIC BEHAVIOR DO 2945 02:16:18,838 --> 02:16:20,206 NOT LEAD TO A RATIONALE OUTCOME 2946 02:16:20,206 --> 02:16:22,208 FOR THE OUTCOME OF SOCIETY. 2947 02:16:22,208 --> 02:16:23,677 SIMPLY PUT, WE CANNOT ALWAYS 2948 02:16:23,677 --> 02:16:26,179 RELY ON CAPITALISM TO GET THE 2949 02:16:26,179 --> 02:16:28,715 RIGHT OUTCOME WHEN IT COMES TO 2950 02:16:28,715 --> 02:16:29,516 THE PUBLIC GOOD. 2951 02:16:29,516 --> 02:16:31,451 FOR BETTER PANDEMIC 2952 02:16:31,451 --> 02:16:33,186 PREPAREDNESS, PFIZER COULD HAVE 2953 02:16:33,186 --> 02:16:35,622 BEEN INCENTIVIZED WITH PUBLIC 2954 02:16:35,622 --> 02:16:38,325 MONEY IN THE 2000s TO CONTINUE 2955 02:16:38,325 --> 02:16:41,828 WORKING ON DISCOVERY OF BROAD 2956 02:16:41,828 --> 02:16:43,963 SPECTRUM M PRO INHIBITORS. 2957 02:16:43,963 --> 02:16:45,532 WE KNOW INVESTMENT OF A FEW 2958 02:16:45,532 --> 02:16:48,868 HUNDRED MILLION U.S. DOLLARS IN 2959 02:16:48,868 --> 02:16:53,506 A SARS COV1 INHIBITOR COULD HAVE 2960 02:16:53,506 --> 02:16:56,142 SAVED MILLIONS BEFORE THE mRNA 2961 02:16:56,142 --> 02:16:57,510 VACCINES BECAME AVAILABLE AND 2962 02:16:57,510 --> 02:17:00,714 PREVENTED ECONOMIC LOSSES TO THE 2963 02:17:00,714 --> 02:17:05,985 TRILLIONS OF U.S. DOLLARS. 2964 02:17:05,985 --> 02:17:08,888 WHEN THE MERS EPIDEMIC STRUCK 2965 02:17:08,888 --> 02:17:11,424 THERE SHOULD HAVE BEEN THE 2966 02:17:11,424 --> 02:17:14,060 DESIRE ABILITY OF MEASURES 2967 02:17:14,060 --> 02:17:16,663 TARGET BEING NEW CORONAVIRUSES. 2968 02:17:16,663 --> 02:17:19,733 EXPERTS KNEW IT WAS NO LONGER OF 2969 02:17:19,733 --> 02:17:22,769 IF BUT WHEN THE NEXT CORONAVIRUS 2970 02:17:22,769 --> 02:17:23,803 EPIDEMIC WOULD STRIKE WITH THE 2971 02:17:23,803 --> 02:17:26,673 RISK OF BECOMING A GLOBAL 2972 02:17:26,673 --> 02:17:27,173 PANDEMIC. 2973 02:17:27,173 --> 02:17:31,611 THANKS TO THE STRUCTURAL BIOLOGY 2974 02:17:31,611 --> 02:17:34,848 COMMUNITY AND OPEN ACCESS PDB 2975 02:17:34,848 --> 02:17:36,750 EXPERTS KNEW THE ACTIVE SITES 2976 02:17:36,750 --> 02:17:38,818 WERE NEARLY IDENTICAL IN 3-D 2977 02:17:38,818 --> 02:17:41,521 STRUCTURE AND THEY'RE POLY 2978 02:17:41,521 --> 02:17:45,191 PROTEIN CUT SITE AMINO ACID 2979 02:17:45,191 --> 02:17:48,595 SEQUENCES WERE SIMILAR. 2980 02:17:48,595 --> 02:17:51,931 THE FREE MARKET FAILED BUT AND 2981 02:17:51,931 --> 02:17:53,299 THE DATA BANK STORED THE 2982 02:17:53,299 --> 02:17:53,600 INFORMATION. 2983 02:17:53,600 --> 02:17:57,170 WE HAVE PFIZER MAKING SIZABLE 2984 02:17:57,170 --> 02:18:00,073 INVESTMENT IN DEVELOPMENT OF 2985 02:18:00,073 --> 02:18:01,508 SARS COV1M PRO INHIBITOR SETTING 2986 02:18:01,508 --> 02:18:04,444 THE STAGE FOR RAPID DISCOVERY 2987 02:18:04,444 --> 02:18:05,812 AND DEVELOPMENT AND EMERGENCY 2988 02:18:05,812 --> 02:18:12,318 USE AUTHORIZATION OF PAXLOVID. 2989 02:18:12,318 --> 02:18:13,119 SIMILARLY, THE FEDERAL 2990 02:18:13,119 --> 02:18:15,188 GOVERNMENT AND PFIZER INVESTED 2991 02:18:15,188 --> 02:18:17,457 IN VACCINE TECHNOLOGY. 2992 02:18:17,457 --> 02:18:21,194 ONE OF THE FEW SILVER LINING IS 2993 02:18:21,194 --> 02:18:22,629 VACCINE DISCOVERY AND 2994 02:18:22,629 --> 02:18:24,397 DEVELOPMENT HAS BEEN 2995 02:18:24,397 --> 02:18:25,498 REVOLUTIONIZED. 2996 02:18:25,498 --> 02:18:30,303 OVER THE NEXT THREE YEARS WE CAN 2997 02:18:30,303 --> 02:18:32,972 SEE HOPE TO PATIENTS AND 2998 02:18:32,972 --> 02:18:35,842 FAMILIES IN THE AREA OF CANCER. 2999 02:18:35,842 --> 02:18:41,815 I'M ALSO HOPEFUL A MORE 3000 02:18:41,815 --> 02:18:44,517 CONCERTED AND EFFECTIVE APPROACH 3001 02:18:44,517 --> 02:18:46,352 TO PANDEMIC PREPAREDNESS 3002 02:18:46,352 --> 02:18:46,686 WORLDWIDE. 3003 02:18:46,686 --> 02:18:48,254 I'VE EXPLAINED HOW OPEN ACCESS 3004 02:18:48,254 --> 02:18:50,523 TO RESEARCH DATA GENERATED WITH 3005 02:18:50,523 --> 02:18:52,025 PUBLIC AND PRIVATE FUNDS 3006 02:18:52,025 --> 02:18:55,628 PARTICULARLY THOUSANDS OF PDB 3007 02:18:55,628 --> 02:18:57,497 STRUCTURES ENABLED BASIC AND 3008 02:18:57,497 --> 02:19:00,733 APPLIES RESEARCHERS TO COMBAT 3009 02:19:00,733 --> 02:19:02,936 COVID-19 PANDEMIC TO SAVE TENS 3010 02:19:02,936 --> 02:19:04,737 OF MILLIONS OF LIVES. 3011 02:19:04,737 --> 02:19:08,074 JOIN ME AND DR. FAUCI IN LOVING 3012 02:19:08,074 --> 02:19:08,708 THE DATA BANK. 3013 02:19:08,708 --> 02:19:09,676 THANK YOU FOR YOUR ATTENTION AND 3014 02:19:09,676 --> 02:19:13,513 ALL THE PEOPLE DEPICTED IN THE 3015 02:19:13,513 --> 02:19:17,250 SLIDE AND OUR WORLDWIDE PROTEIN 3016 02:19:17,250 --> 02:19:18,251 DATA BANK COLLEAGUES WHO MADE 3017 02:19:18,251 --> 02:19:20,820 THIS POSSIBLE, OUR FUNDERS AND 3018 02:19:20,820 --> 02:19:26,559 NSF AND NIH AND WE'RE LISTED 3019 02:19:26,559 --> 02:19:29,162 HERE AS ARE OUR HOSTS AT RUTGERS 3020 02:19:29,162 --> 02:19:31,698 AND UCSD AND UCSF. 3021 02:19:31,698 --> 02:19:33,066 THANK YOU FOR YOUR ATTENTION. 3022 02:19:33,066 --> 02:19:35,068 >> THANK YOU FOR A WONDERFUL 3023 02:19:35,068 --> 02:19:35,268 TALK. 3024 02:19:35,268 --> 02:19:36,803 IF YOU CAN TURN YOUR CAMERA ON 3025 02:19:36,803 --> 02:19:39,005 TO TAKE A QUESTION OR TWO. 3026 02:19:39,005 --> 02:19:42,308 LET'S TAKE ONE OR TWO QUESTIONS. 3027 02:19:42,308 --> 02:19:45,111 WE CAN'T TAKE MORE BECAUSE WE'RE 3028 02:19:45,111 --> 02:19:46,179 BLEEDING INTO THIS BREAK. 3029 02:19:46,179 --> 02:19:47,347 IS THERE PUBLICATION THAT 3030 02:19:47,347 --> 02:19:49,549 DETAILS THE PROTOCOLS OR 3031 02:19:49,549 --> 02:19:54,687 PROFESSIONS OF PDB CURATION? 3032 02:19:54,687 --> 02:20:01,561 >> THERE IS A JOURNAL BIO 3033 02:20:01,561 --> 02:20:04,998 CURATION THAT DOCUMENTS ALL OF 3034 02:20:04,998 --> 02:20:05,398 THAT. 3035 02:20:05,398 --> 02:20:09,068 WE HAVE ON OUR DATA BANK DOT ORG 3036 02:20:09,068 --> 02:20:12,171 WEBSITE A DETAILED EXPLANATION 3037 02:20:12,171 --> 02:20:15,408 OF CURATION PROCESSES UNDERTAKEN 3038 02:20:15,408 --> 02:20:17,510 WHEN EVERY NEW STRUCTURE COMES 3039 02:20:17,510 --> 02:20:21,514 IN THE PROTEIN DATA BANK. 3040 02:20:21,514 --> 02:20:24,851 THE PROTOCOLS FOR VALIDATION AND 3041 02:20:24,851 --> 02:20:27,887 BIO CURATION ARE GLOBALLY 3042 02:20:27,887 --> 02:20:30,256 UNIFORM BETWEEN COLLEAGUES IN 3043 02:20:30,256 --> 02:20:32,325 JAPAN AND THE PEOPLES REPUBLIC 3044 02:20:32,325 --> 02:20:34,060 OF CHINA, EUROPE AND THE U.S. 3045 02:20:34,060 --> 02:20:35,428 >> THANK YOU. 3046 02:20:35,428 --> 02:20:37,630 NEXT QUESTION, HOW DO THE 3047 02:20:37,630 --> 02:20:39,232 DETERMINE ACTIVE SITES OF THE 3048 02:20:39,232 --> 02:20:42,902 PROTEIN WITH NO REFERENCE 3049 02:20:42,902 --> 02:20:43,937 STRUCTURES OR LIGANDS. 3050 02:20:43,937 --> 02:20:47,640 IS THERE A WEBSITE TO HELP WITH 3051 02:20:47,640 --> 02:20:47,907 THIS? 3052 02:20:47,907 --> 02:20:52,645 >> A TEAM DID AN ANALYSIS OF 3053 02:20:52,645 --> 02:20:53,913 ENZYME ACTIVE SITES MANY YEARS 3054 02:20:53,913 --> 02:20:54,080 AGO. 3055 02:20:54,080 --> 02:20:56,416 I DON'T RECALL WHERE IT WAS 3056 02:20:56,416 --> 02:21:01,254 PUBLISHED AND THE THRUST OF 3057 02:21:01,254 --> 02:21:03,456 THEIR RESULTS WAS LIKELY TO BE 3058 02:21:03,456 --> 02:21:07,126 THE ENZYME ACTIVE SITE WHEN 3059 02:21:07,126 --> 02:21:10,730 DEALING WITH A CATALYTIC ENTITY. 3060 02:21:10,730 --> 02:21:12,932 >> FANTASTIC. 3061 02:21:12,932 --> 02:21:13,499 THANK YOU. 3062 02:21:13,499 --> 02:21:17,503 NEXT QUESTION, WE'LL TAKE ONE 3063 02:21:17,503 --> 02:21:18,771 MORE. 3064 02:21:18,771 --> 02:21:22,275 RCSB OFFERS 3-D STRUCTURES IN 3065 02:21:22,275 --> 02:21:25,511 PDB 5 FORMAT WITH OTHERS. 3066 02:21:25,511 --> 02:21:27,914 MANY RESEARCHERS FAVORS THE 3067 02:21:27,914 --> 02:21:30,249 FORMAT BUT IT HAS LIMITATIONS. 3068 02:21:30,249 --> 02:21:33,186 ARE THERE PLANS TO MODIFY THE 3069 02:21:33,186 --> 02:21:35,555 FORMAT OR A NEW ONE TO INCLUDE 3070 02:21:35,555 --> 02:21:36,856 MORE DETAILED INFORMATION ABOUT 3071 02:21:36,856 --> 02:21:40,293 PROTEINS IN A USER FRIENDLY WAY? 3072 02:21:40,293 --> 02:21:41,894 >> EXCELLENT QUESTION. 3073 02:21:41,894 --> 02:21:43,896 THAT'S ALREADY HAPPENED. 3074 02:21:43,896 --> 02:21:48,167 THE ARCHIVAL FORMAT FOR THE PDB 3075 02:21:48,167 --> 02:21:51,170 IS THE PDB DATA STANDARD CLOSELY 3076 02:21:51,170 --> 02:21:53,506 RELATED TO THE DERIVATIVE 3077 02:21:53,506 --> 02:21:59,212 STANDARD USED BY ALPHA FOLD DB 3078 02:21:59,212 --> 02:22:02,515 AND THE MODEL ARCHIVE THE 3079 02:22:02,515 --> 02:22:05,251 ADVANTAGE IS IT'S BOTH MACHINE 3080 02:22:05,251 --> 02:22:08,688 AND HUMAN READABLE. 3081 02:22:08,688 --> 02:22:12,725 IT'S INFINITELY EXTENSIBLE TO 3082 02:22:12,725 --> 02:22:13,493 STRUCTURE SIZE AND COMPLEXITY 3083 02:22:13,493 --> 02:22:17,497 AND THE ADDITION OF NEW DATA 3084 02:22:17,497 --> 02:22:21,167 STANDARDS TO SUPPORT NEW TYPES 3085 02:22:21,167 --> 02:22:24,237 OF EXPERIMENT AND THINGS LIKE 3086 02:22:24,237 --> 02:22:25,772 POST TRANSLATIONAL 3087 02:22:25,772 --> 02:22:30,910 MODIFICATIONS, ETCETERA. 3088 02:22:30,910 --> 02:22:35,815 I KNOW THE 80 COLUMN PUNCH CARD 3089 02:22:35,815 --> 02:22:37,750 FORMAT OF THE ORIGINAL PDB 3090 02:22:37,750 --> 02:22:39,852 FORMAT IS VERY EASY FOR PEOPLE 3091 02:22:39,852 --> 02:22:43,322 TO USE BECAUSE THEY'RE FAMILIAR. 3092 02:22:43,322 --> 02:22:45,191 BUT I THINK IN THE LONG RUN WE 3093 02:22:45,191 --> 02:22:47,627 ALL HAVE TO MOVE TO THE DATA 3094 02:22:47,627 --> 02:22:50,463 STANDARD IF WE WANT THE PDB TO 3095 02:22:50,463 --> 02:22:54,901 BE USEFUL GOING FORWARD. 3096 02:22:54,901 --> 02:22:57,170 MANY STRUCTURES WE'RE NOW 3097 02:22:57,170 --> 02:23:01,340 ARCHIVING CANNOT BE DEPICTED IN 3098 02:23:01,340 --> 02:23:04,043 THE LEGACY FORMAT. 3099 02:23:04,043 --> 02:23:05,511 THOSE STRUCTURED DATA FILES ARE 3100 02:23:05,511 --> 02:23:07,747 NO LONGER AVAILABLE IN LEGACY 3101 02:23:07,747 --> 02:23:11,050 PDB FORMAT. 3102 02:23:11,050 --> 02:23:15,188 THEY'RE ONLY AVAILABLE IN THE 3103 02:23:15,188 --> 02:23:16,122 DATA STANDARD. 3104 02:23:16,122 --> 02:23:21,194 THOSE CONSUMING PDB DATA GO WITH 3105 02:23:21,194 --> 02:23:22,829 THE FLOW AND YOU'LL FIND THE 3106 02:23:22,829 --> 02:23:25,398 STANDARD IS EASY TO LEARN, EASY 3107 02:23:25,398 --> 02:23:27,333 TO READ AND EASY TO, WOULD WITH 3108 02:23:27,333 --> 02:23:29,502 FROM A SOFTWARE ENGINEERING 3109 02:23:29,502 --> 02:23:33,172 STANDPOINT. 3110 02:23:33,172 --> 02:23:34,207 >> FANTASTIC THANK YOU FOR A 3111 02:23:34,207 --> 02:23:34,807 WONDERFUL PRESENTATION AND 3112 02:23:34,807 --> 02:23:37,110 TAKING ON THE QUESTIONS. 3113 02:23:37,110 --> 02:23:40,246 WE APPRECIATE IT SO MUCH. 3114 02:23:40,246 --> 02:23:42,014 ALL RIGHT, EVERYONE. 3115 02:23:42,014 --> 02:23:48,087 I'LL SHARE MY SLIDE DECK SO NEXT 3116 02:23:48,087 --> 02:23:53,826 WE'LL HAVE A SHORT BREAK AND 3117 02:23:53,826 --> 02:23:58,364 WE'LL START SESSION 2 AT 2:00. 3118 02:23:58,364 --> 02:24:01,501 10 MINUTE BREAK AND BACK ON IN 3119 02:24:01,501 --> 02:24:03,269 SILICO METHODOLOGIES USED IN 3120 02:24:03,269 --> 02:24:04,135 DRUG DISCOVERY. SEE YOU IN 01 MINUTES. 3121 02:24:06,316 --> 02:24:08,184 >> THANK YOU WELCOME BACK TO 3122 02:24:08,184 --> 02:24:08,818 SESSION 2. 3123 02:24:08,818 --> 02:24:11,754 SUPER EXCITING SESSION FOR YOU 3124 02:24:11,754 --> 02:24:16,292 ON IN SILICO METHODOLOGIES USED 3125 02:24:16,292 --> 02:24:21,831 IN DRUG DISCOVERY IT'S 3126 02:24:21,831 --> 02:24:29,572 CO-CHAIRED BY ALEXANDRE VARNEK 3127 02:24:29,572 --> 02:24:33,409 AT THE SCHOOL -- ALEXANDER 3128 02:24:33,409 --> 02:24:35,178 TROPSHA FROM THE UNIVERSITY OF 3129 02:24:35,178 --> 02:24:36,245 NORTH CAROLINA CHAPEL HILL AND 3130 02:24:36,245 --> 02:24:39,615 IS AN ELECTED FELLOW OF THE 3131 02:24:39,615 --> 02:24:40,416 AMERICAN INSTITUTE OF 3132 02:24:40,416 --> 02:24:43,286 ENGINEERING AND A CO-FOUNDER OF 3133 02:24:43,286 --> 02:24:47,090 PREDICTIVE LLC A UNC START-UP 3134 02:24:47,090 --> 02:24:51,661 SPECIALIZING IN BUILDING MODELS 3135 02:24:51,661 --> 02:24:54,397 AND TOOLS FOR PREDICTION AND 3136 02:24:54,397 --> 02:24:56,132 HELPED PUT THE WORK SHOP 3137 02:24:56,132 --> 02:24:56,532 TOGETHER. 3138 02:24:56,532 --> 02:24:58,201 THANK YOU, THE FLOOR IS YOURS. 3139 02:24:58,201 --> 02:24:59,569 >> I'LL BE BRIEF. 3140 02:24:59,569 --> 02:25:03,806 I WANT TO START BY THANKING 3141 02:25:03,806 --> 02:25:05,141 SPEAKING FOR HIGHLIGHTING THE 3142 02:25:05,141 --> 02:25:08,211 IMPORTANCE OF DATA AND DATABASES 3143 02:25:08,211 --> 02:25:09,512 THAT CAN BE USED FOR 3144 02:25:09,512 --> 02:25:10,246 APPLICATIONS. 3145 02:25:10,246 --> 02:25:14,450 IN THIS SESSION, WE'LL HEAR FROM 3146 02:25:14,450 --> 02:25:21,290 FOUR EXPERTS IN ANALYTICS, USING 3147 02:25:21,290 --> 02:25:23,860 AVAILABLE AND SYNTHETICALLY 3148 02:25:23,860 --> 02:25:24,627 FEASIBLE METHODS OF MACHINE 3149 02:25:24,627 --> 02:25:27,296 LEARNING AND FEASIBLE MODELLING. 3150 02:25:27,296 --> 02:25:34,370 THE FIRST SPEAKER IS ALEXANDER 3151 02:25:34,370 --> 02:25:39,475 VARNEK OF CHEMISTRY AND THE 3152 02:25:39,475 --> 02:25:42,445 DIRECTOR OF THE SCIENCE AT THE 3153 02:25:42,445 --> 02:25:48,818 UNIVERSITY OF STRASBOURG, FRANCE 3154 02:25:48,818 --> 02:25:51,988 AND PART OF A GROUP AND 3155 02:25:51,988 --> 02:25:54,023 PUBLISHED BOOKS AND ARTICLES AND 3156 02:25:54,023 --> 02:25:57,393 CHAPTERS AND IS AN EDI 3157 02:25:57,393 --> 02:26:02,698 EDITOR-IN-CHIEF AND HAS BEEN 3158 02:26:02,698 --> 02:26:04,534 HIGHLIGHTED IN WORKSHOPS IN THE 3159 02:26:04,534 --> 02:26:06,903 LAST YEARS IN STRASBOURG AND HE 3160 02:26:06,903 --> 02:26:10,873 IS THE RECIPIENT, 3161 02:26:10,873 --> 02:26:12,475 CONGRATULATIONS, AGAIN FOR THE 3162 02:26:12,475 --> 02:26:14,010 AWARD FROM THE AMERICAN CHEMICAL 3163 02:26:14,010 --> 02:26:14,477 SOCIETY. 3164 02:26:14,477 --> 02:26:20,817 YOU HAVE THE FLOOR. 3165 02:26:20,817 --> 02:26:31,227 >> THANK YOU VERY MUCH. 3166 02:26:33,696 --> 02:26:35,398 I'LL SPEAK ABOUT SPECIFIC 3167 02:26:35,398 --> 02:26:38,067 METHODS OF EXPLORATION OF LARGE 3168 02:26:38,067 --> 02:26:42,405 CHEMICAL SPACES. 3169 02:26:42,405 --> 02:26:48,344 NOWADAYS, CHEMISTRY GENERATES A 3170 02:26:48,344 --> 02:26:52,548 LOT OF DATA. 3171 02:26:52,548 --> 02:26:59,689 AND DATABASES AND MORE THAN 26 3172 02:26:59,689 --> 02:27:05,828 ARE STORED IN THE DATABASES OF 3173 02:27:05,828 --> 02:27:07,530 LARGE COMPANIES. 3174 02:27:07,530 --> 02:27:12,068 THE CHEMICAL LIBRARY IS BASED ON 3175 02:27:12,068 --> 02:27:15,071 THE SIZE CORRELATES WITH 3176 02:27:15,071 --> 02:27:16,539 SCREENING. 3177 02:27:16,539 --> 02:27:21,344 THIS HOWEVER IS NOT OBVIOUS. 3178 02:27:21,344 --> 02:27:23,679 THE QUESTION ARISES HOW TO FIND 3179 02:27:23,679 --> 02:27:27,383 THE LIBRARY THAT'S THE MOST 3180 02:27:27,383 --> 02:27:34,557 SUITABLE FOR A PARTICULAR 3181 02:27:34,557 --> 02:27:36,792 MOLECULE. 3182 02:27:36,792 --> 02:27:42,832 I WOULD LIKE TO PRESENT THE 3183 02:27:42,832 --> 02:27:48,804 APPROACH WHICH INCORPORATION OF 3184 02:27:48,804 --> 02:27:52,041 THE LIBRARIES AND ALSO IN 3185 02:27:52,041 --> 02:27:53,976 COMBINATION WITH GROUPS MAY HELP 3186 02:27:53,976 --> 02:28:02,184 DESIGN NOVEL MOLECULAR ENTITIES. 3187 02:28:02,184 --> 02:28:04,487 CHEMOGRAPHY CONSIDERS DATA 3188 02:28:04,487 --> 02:28:06,656 DISTRIBUTION ON MAPS AND 3189 02:28:06,656 --> 02:28:09,492 STRUCTURES AND HOW SUCH MAPS 3190 02:28:09,492 --> 02:28:14,830 COULD BE OBTAINED. 3191 02:28:14,830 --> 02:28:17,466 A MOLECULE CAN BE PRESENTED AS 3192 02:28:17,466 --> 02:28:19,435 AN OBJECT IN THE CHEMICAL SPACE 3193 02:28:19,435 --> 02:28:25,341 IN ORDER TO OBTAIN THE MAP WE 3194 02:28:25,341 --> 02:28:35,418 NEED TO USE THIS. 3195 02:28:35,418 --> 02:28:39,388 THERE'S VARIOUS METHODS AND IN 3196 02:28:39,388 --> 02:28:42,124 TERMS OF DATA DISTRIBUTION. 3197 02:28:42,124 --> 02:28:48,798 BUT SOME REASON WE SELECTED THE 3198 02:28:48,798 --> 02:28:58,074 METHOD THIS SLIDE SHOWS THE 3199 02:28:58,074 --> 02:28:59,108 ALGORITHM. 3200 02:28:59,108 --> 02:29:02,144 FROM THE LARGE CHEMICAL LIBRARY 3201 02:29:02,144 --> 02:29:07,350 WE SELECT A SMALL REPRESENTATIVE 3202 02:29:07,350 --> 02:29:17,893 DATA SET AND THEN THE ALGORITHM 3203 02:29:19,328 --> 02:29:20,796 MOLECULE. 3204 02:29:20,796 --> 02:29:31,340 THIS IS POSITION APPROACHES THE 3205 02:29:40,616 --> 02:29:41,317 MAP. 3206 02:29:41,317 --> 02:29:44,487 THIS PROJECTION IS DESCRIBED BY 3207 02:29:44,487 --> 02:29:53,896 THE LOCATION IN DIFFERENT MODES. 3208 02:29:53,896 --> 02:30:00,469 SUCH PROJECTION IS ACROSS AND 3209 02:30:00,469 --> 02:30:03,406 HAS DISTRIBUTION AS YOU MAY SEE 3210 02:30:03,406 --> 02:30:03,973 IN THE RIGHT CORNER OF THIS 3211 02:30:03,973 --> 02:30:11,080 SLIDE. 3212 02:30:11,080 --> 02:30:16,786 IN TURN, THIS DISTRIBUTION CAN 3213 02:30:16,786 --> 02:30:27,129 BE INCLUDED RELATED. 3214 02:30:27,496 --> 02:30:29,532 AND WE HAVE INDIVIDUAL OBJECTS 3215 02:30:29,532 --> 02:30:34,103 AND IT CAN CONSIDER THE 3216 02:30:34,103 --> 02:30:35,571 PROBABILITY DISTRIBUTION AND 3217 02:30:35,571 --> 02:30:41,277 OPENS AN OPPORTUNITY FOR THE 3218 02:30:41,277 --> 02:30:49,585 APPROACH FOR MODERN TASKS. 3219 02:30:49,585 --> 02:30:52,555 AND THERE'S INFORMATION ABOUT 3220 02:30:52,555 --> 02:30:53,856 VARIOUS MOLECULAR PROPERTIES 3221 02:30:53,856 --> 02:30:56,125 USING GTM LANDSCAPES. 3222 02:30:56,125 --> 02:30:58,527 AND ONE CAN CONSIDER THREE TYPES 3223 02:30:58,527 --> 02:31:00,629 OF LANDSCAPES. 3224 02:31:00,629 --> 02:31:02,732 DENSE LANDSCAPE, CHARACTERIZING 3225 02:31:02,732 --> 02:31:06,302 THE DATA DENSITY, THE CLASS 3226 02:31:06,302 --> 02:31:10,573 LANDSCAPE AND THE POPULATION IN 3227 02:31:10,573 --> 02:31:16,545 DIFFERENT AREAS OF THE MAP AND 3228 02:31:16,545 --> 02:31:20,616 ACTIVITY LANDSCAPE. 3229 02:31:20,616 --> 02:31:22,551 EACH OF THESE MAPS CAN BE 3230 02:31:22,551 --> 02:31:24,186 ENCODED BY THE DISTRIBUTION OF 3231 02:31:24,186 --> 02:31:28,557 CHEMO TYPE AND ACTIVITIES IN THE 3232 02:31:28,557 --> 02:31:32,628 CHEMICAL SPACE. 3233 02:31:32,628 --> 02:31:34,163 THE LANDSCAPE CHARACTERIZES THE 3234 02:31:34,163 --> 02:31:36,298 DATA IN THE DISTRIBUTION. 3235 02:31:36,298 --> 02:31:41,270 THEY CAN BE USEFUL. 3236 02:31:41,270 --> 02:31:44,840 THUS FOR THE MODE THE LANDSCAPE 3237 02:31:44,840 --> 02:31:50,446 CAN BE USED TO PREDICT A CLASS 3238 02:31:50,446 --> 02:31:52,581 AND WORKS AS A CLASSIFICATION 3239 02:31:52,581 --> 02:31:58,087 MODEL. 3240 02:31:58,087 --> 02:31:59,922 THE LANDSCAPE WORKS LIKE A MODEL 3241 02:31:59,922 --> 02:32:04,794 TO PREDICT THE ACTIVITY AND IF 3242 02:32:04,794 --> 02:32:07,329 THE COMPOUND IS PROJECTED IN 3243 02:32:07,329 --> 02:32:10,966 THIS AREA IT'S CONSIDERED TO BE 3244 02:32:10,966 --> 02:32:14,236 OUT OF THE MODEL OF THE DOMAIN. 3245 02:32:14,236 --> 02:32:16,605 IT CAN BE USED TO ANALYZE 3246 02:32:16,605 --> 02:32:19,308 CHEMICAL SPACE AND LIBRARIES. 3247 02:32:19,308 --> 02:32:26,048 HERE WE CONSIDER THE CHEMBL 3248 02:32:26,048 --> 02:32:27,216 DATABASE CONTAINING 2 MILLION 3249 02:32:27,216 --> 02:32:30,219 PROPERTIES AND ACTIVITIES. 3250 02:32:30,219 --> 02:32:34,557 6:ON THE NEXT ONE CAN EASILY 3251 02:32:34,557 --> 02:32:40,029 IDENTIFY FIVE DENSITY ZONES 3252 02:32:40,029 --> 02:32:48,804 WHICH CAN BE ASSOCIATED. 3253 02:32:48,804 --> 02:32:51,807 BIOLOGICAL ACTIVITIES ARE SHOWN 3254 02:32:51,807 --> 02:32:52,575 ON THE SLIDE. 3255 02:32:52,575 --> 02:32:56,579 EACH LANDSCAPE IS STUDIED FOR 3256 02:32:56,579 --> 02:32:58,581 BIOLOGICAL TARGETS. 3257 02:32:58,581 --> 02:33:02,251 WE BUILT MORE THAN 700 OF DATA 3258 02:33:02,251 --> 02:33:03,219 AND FOR ANY NEW COMPOUND 3259 02:33:03,219 --> 02:33:05,254 PROJECTED ON THE MAP ONE CAN 3260 02:33:05,254 --> 02:33:08,791 IDENTIFY THE MOST SIMILAR 3261 02:33:08,791 --> 02:33:10,993 COMPOUND. 3262 02:33:10,993 --> 02:33:14,663 AND PREDICT THE PHARMACOLOGIC 3263 02:33:14,663 --> 02:33:20,769 PROFILE OF ACTIVITIES USING THIS 3264 02:33:20,769 --> 02:33:21,270 ACTIVITY OF LANDSCAPES. 3265 02:33:21,270 --> 02:33:25,875 THE NEXT PART OF MY TALK HAS 3266 02:33:25,875 --> 02:33:32,815 PAIRWISE COMPARISON FOR A GIVEN 3267 02:33:32,815 --> 02:33:34,884 LIBRARY. 3268 02:33:34,884 --> 02:33:37,253 I'LL SHOW A FEW EXAMPLES AND THE 3269 02:33:37,253 --> 02:33:44,760 FIRST HAS PAIR WISE COMPARISON 3270 02:33:44,760 --> 02:33:52,568 FOR COMMERCIALLY AVAILABLE 3271 02:33:52,568 --> 02:34:00,042 COMPOUNDS ALL TOGETHER 3272 02:34:00,042 --> 02:34:03,312 STRUCTURES WERE CONSIDERED AND 3273 02:34:03,312 --> 02:34:07,750 WE TESTED AND WHICH WOULD 3274 02:34:07,750 --> 02:34:10,552 ENHANCE THE LIBRARIES. 3275 02:34:10,552 --> 02:34:12,821 ON THE OTHER HAND CHEMBL COULD 3276 02:34:12,821 --> 02:34:16,558 BE RECOMMENDED FOR THE 3277 02:34:16,558 --> 02:34:26,802 BIOLOGICAL MEASURES. 3278 02:34:26,802 --> 02:34:29,805 WE HAD DISCOVERED DIFFERENT 3279 02:34:29,805 --> 02:34:31,173 COMPOUND CLASSES. 3280 02:34:31,173 --> 02:34:39,815 AND THIS ACCOMMODATES THESE AND 3281 02:34:39,815 --> 02:34:43,953 WE NEEDED TO IDENTIFY THE ZONE 3282 02:34:43,953 --> 02:34:49,325 OR COMPOUNDS FROM ONE DATABASE. 3283 02:34:49,325 --> 02:34:59,802 THIS PART WAS NOT ABLE TO 3284 02:35:04,640 --> 02:35:08,811 IDENTIFY THE EXISTING LARGER 3285 02:35:08,811 --> 02:35:16,552 SPECIFIC SOURCE AND FOR EACH 3286 02:35:16,552 --> 02:35:26,962 SELECTED ZONE WE MAPPED. 3287 02:35:28,597 --> 02:35:39,074 AND WE CAN CONTINUE THE MAPS AND 3288 02:35:39,541 --> 02:35:48,917 THIS CONTAINED COMPOUNDS AND 3289 02:35:48,917 --> 02:35:51,587 POPULATED BY ZINC OR THE 3290 02:35:51,587 --> 02:35:53,822 COMPOUNDS FROM WHICH WE CAN 3291 02:35:53,822 --> 02:35:59,428 CONSTRUCT PARTICULAR CHEMO TYPES 3292 02:35:59,428 --> 02:36:04,600 AND STRUCTURES. 3293 02:36:04,600 --> 02:36:08,137 AND THIS ALLOWS US TO EXTRACT 3294 02:36:08,137 --> 02:36:11,607 MORE THAN 1,000 TYPES NOT BIO 3295 02:36:11,607 --> 02:36:14,610 LOGICALLY TESTED AND MORE THAN 3296 02:36:14,610 --> 02:36:18,914 20,000 CHEMO TYPES IN THE 3297 02:36:18,914 --> 02:36:25,788 COMMERCIAL CHEMICAL SPACE AND 3298 02:36:25,788 --> 02:36:26,822 MITIGATED THE CHEMICAL SPACE 3299 02:36:26,822 --> 02:36:33,662 WITH SO MANY INDIVIDUAL MAPS. 3300 02:36:33,662 --> 02:36:42,204 SO NEXT EXAMPLE AND THIS 3301 02:36:42,204 --> 02:36:44,406 COLLECTION CONTAINS 2.2 MILLION 3302 02:36:44,406 --> 02:36:48,811 COMPOUNDS. 3303 02:36:48,811 --> 02:36:53,515 THE COMPANY INTENDED TO 3304 02:36:53,515 --> 02:37:00,756 DIVERSIFY THE LIBRARY FROM THE 3305 02:37:00,756 --> 02:37:01,857 DATABASE CONTAINING MORE THAN 3306 02:37:01,857 --> 02:37:06,695 WHAT WAS LIMITED. 3307 02:37:06,695 --> 02:37:15,671 SO WE PROPOSED THE UNIQUE 3308 02:37:15,671 --> 02:37:20,809 MAXIMUM MANY STRATEGIC TOURS 3309 02:37:20,809 --> 02:37:29,985 FROM SEGMA AND ALSO USED THE 3310 02:37:29,985 --> 02:37:37,960 ACTIVITY AND IN SUCH A WAY WE 3311 02:37:37,960 --> 02:37:44,833 SELECTED SOME COMPOUNDS BEFORE 3312 02:37:44,833 --> 02:37:52,908 THE PROPERTY SELECTION. 3313 02:37:52,908 --> 02:37:56,612 SO THIS DELIVERED SPACE. 3314 02:37:56,612 --> 02:38:00,816 THIS IS FROM A LIBRARY AND IN 3315 02:38:00,816 --> 02:38:03,318 THIS DEMONSTRATED THE SPACE IS 3316 02:38:03,318 --> 02:38:09,625 NOT REDUNDANT BUT COMPLIMENTARY 3317 02:38:09,625 --> 02:38:13,896 CONTAINING 40 BILLION STRUCTURES 3318 02:38:13,896 --> 02:38:18,700 AND THIS IS REPRESENTED A SUBSET 3319 02:38:18,700 --> 02:38:23,105 CHOSE LITTLE IN THE SPACES. 3320 02:38:23,105 --> 02:38:28,644 SO YOU CAN SEE HERE BETWEEN HERE 3321 02:38:28,644 --> 02:38:35,384 AND WHEN YOU LOOK AT THE 3322 02:38:35,384 --> 02:38:40,422 LANDSCAPES WE CAN SAY THE 3323 02:38:40,422 --> 02:38:50,933 DATABASE IS ENRICHED BY MORE 3324 02:38:52,634 --> 02:38:56,939 INDEXES IN THE COMPOUNDS. 3325 02:38:56,939 --> 02:38:58,240 OKAY. 3326 02:38:58,240 --> 02:39:02,844 LET'S NOW CONSIDER MULTIPLE 3327 02:39:02,844 --> 02:39:03,111 ANALYSIS. 3328 02:39:03,111 --> 02:39:08,083 HERE WE COMPARED ONE LIBRARY AND 3329 02:39:08,083 --> 02:39:10,919 THE GOAL IS TO SELECT LIBRARIES 3330 02:39:10,919 --> 02:39:12,788 THAT RESPOND ACCORDING TO THE 3331 02:39:12,788 --> 02:39:13,322 SCORE. 3332 02:39:13,322 --> 02:39:17,326 AND IN THIS STUDY WE CONSIDERED 3333 02:39:17,326 --> 02:39:19,428 THE NAME FOR THE LIBRARY AS 3334 02:39:19,428 --> 02:39:21,430 EXTERNAL AND CHEMBL AS A 3335 02:39:21,430 --> 02:39:22,564 REFERENCE DATABASE. 3336 02:39:22,564 --> 02:39:24,833 SO JUST TO REMIND YOU IT 3337 02:39:24,833 --> 02:39:28,570 REPRESENTS THE COLLECTION OF 3338 02:39:28,570 --> 02:39:36,778 SMALL MOLECULES OR DNA. 3339 02:39:36,778 --> 02:39:40,582 AND USING COLLECTION OF 3340 02:39:40,582 --> 02:39:41,817 COMMERCIALLY AVAILABLE RULES, 3341 02:39:41,817 --> 02:39:46,688 ONE CAN GENERATE THOUSANDS 3342 02:39:46,688 --> 02:39:51,393 CONTAINING BILLIONS AND IT 3343 02:39:51,393 --> 02:39:55,197 SHOULD BE NOTED THAT SYNTHESIS 3344 02:39:55,197 --> 02:39:58,333 IS PRETTY EXPENSIVE. 3345 02:39:58,333 --> 02:40:02,137 THEREFORE SELECTION OF THE 3346 02:40:02,137 --> 02:40:04,039 PARTICULAR DRUG DISCOVERY 3347 02:40:04,039 --> 02:40:04,573 CAMPAIGN IS AN IMPORTANT 3348 02:40:04,573 --> 02:40:14,816 CHEMICAL TASK. 3349 02:40:15,450 --> 02:40:18,820 USING THE LIBRARY CAN REPRESENT 3350 02:40:18,820 --> 02:40:23,292 AND WE CAN EASILY SUGGEST A 3351 02:40:23,292 --> 02:40:27,095 METRIC FOR EXAMPLE A MODEL WHICH 3352 02:40:27,095 --> 02:40:32,801 CAN ASSESS OVERLAP BETWEEN THE 3353 02:40:32,801 --> 02:40:33,402 TWO DATABASES. 3354 02:40:33,402 --> 02:40:36,104 AND THEN PERFORM COMPUTATIONAL 3355 02:40:36,104 --> 02:40:42,377 EXPERIMENT USING A COLLECTION OF 3356 02:40:42,377 --> 02:40:47,549 39,000 MODELS IN THE DATABASE 3357 02:40:47,549 --> 02:40:55,757 AND DESIGNER TOOL AND WE 3358 02:40:55,757 --> 02:40:58,560 DESIGNED THE SIZE FROM EACH OF 3359 02:40:58,560 --> 02:41:02,597 THESE WE SELECT A REPRESENTATIVE 3360 02:41:02,597 --> 02:41:06,401 DATA SET OF 1 MILLION COMPOUND 3361 02:41:06,401 --> 02:41:08,837 AND CONSIDERED 2.5 BILLION 3362 02:41:08,837 --> 02:41:15,510 COMPOUNDS ALTOGETHER AND INSECT 3363 02:41:15,510 --> 02:41:16,712 DELL WITH CHEMBL AND SELECT THE 3364 02:41:16,712 --> 02:41:26,822 SCORE. 3365 02:41:29,558 --> 02:41:33,428 WE DISCOVERED IT ALREADY COVERED 3366 02:41:33,428 --> 02:41:39,768 MORE THAN 50% OF TYPES OF CHEMBL 3367 02:41:39,768 --> 02:41:46,575 AND THEN WE WERE LOOKING FOR THE 3368 02:41:46,575 --> 02:41:57,119 DEL WHICH RECOVERS THE AREA AND 3369 02:41:59,788 --> 02:42:04,559 WE DISCOVERED THREE COVERED MORE 3370 02:42:04,559 --> 02:42:09,765 THAN 80% OF THE SPACE AND 3371 02:42:09,765 --> 02:42:14,870 PROJECTED ON THE MAP WE 3372 02:42:14,870 --> 02:42:20,308 DISCOVERED THE AREA OF COMPOUNDS 3373 02:42:20,308 --> 02:42:29,050 WHICH ARE NOT BY THE DEL 3374 02:42:29,050 --> 02:42:29,851 CHEMISTRY. 3375 02:42:29,851 --> 02:42:35,757 WHEN YOU CONSIDER THE CHEMICAL 3376 02:42:35,757 --> 02:42:44,199 LIBRARIES, WE KCAN CONSIDER THE 3377 02:42:44,199 --> 02:42:45,667 SPACING IN WHICH EACH DATA POINT 3378 02:42:45,667 --> 02:42:52,407 IS AN INDIVIDUAL LIBRARY. 3379 02:42:52,407 --> 02:42:56,778 AND WE CAN VIEW ANOTHER WHICH WE 3380 02:42:56,778 --> 02:43:01,249 CALL GTM AND LOOK AT THE 3381 02:43:01,249 --> 02:43:04,586 RELATIONSHIP BETWEEN LIBRARIES 3382 02:43:04,586 --> 02:43:05,754 IN THIS SPACE. 3383 02:43:05,754 --> 02:43:09,925 THIS IS AN EXAMPLE SO IN WHICH 3384 02:43:09,925 --> 02:43:18,133 WE PROMOTE THE DELS AND 3385 02:43:18,133 --> 02:43:23,238 RESPECTIVE CHEMBL ON THE BOTTOM 3386 02:43:23,238 --> 02:43:33,782 LEFT -- AND THE CLOSEST IN THE 3387 02:43:40,622 --> 02:43:51,099 DATABASE YOU CAN SEE THE CLOSEST 3388 02:43:51,500 --> 02:43:57,606 DEL IS THE MOST SIMILAR TO THE 3389 02:43:57,606 --> 02:44:04,579 LANDSCAPE OF CHEMBL AND CAN 3390 02:44:04,579 --> 02:44:11,119 ANALYZE REACTION USED TO PULLED 3391 02:44:11,119 --> 02:44:20,795 THE DELS AND CONCLUDE THEY HAVE 3392 02:44:20,795 --> 02:44:31,239 FEW DELS AND REACTIONS. 3393 02:44:31,239 --> 02:44:33,942 THEN AND WE HAVE THE APPLICATION 3394 02:44:33,942 --> 02:44:38,346 IN COMBINATION TO GENERATE NOVEL 3395 02:44:38,346 --> 02:44:40,448 MOLECULE STRUCTURES AND CHEMICAL 3396 02:44:40,448 --> 02:44:40,715 REACTIONS. 3397 02:44:40,715 --> 02:44:46,588 SO THE IDEA IS SIMPLE. 3398 02:44:46,588 --> 02:44:57,132 SO LET'S CONSIDER WE CAN USE THE 3399 02:44:58,667 --> 02:45:04,739 11 VECTORS AND CAN LOOK AT WHERE 3400 02:45:04,739 --> 02:45:08,810 WE CAN MANUALLY SELECT AN AREA 3401 02:45:08,810 --> 02:45:15,016 POPULATED BY USEFUL STRUCTURES. 3402 02:45:15,016 --> 02:45:20,822 AND WE APPLIED THIS APPROACH 3403 02:45:20,822 --> 02:45:31,333 SEVERAL YEARS AGO AND WE HAVE 3404 02:45:33,234 --> 02:45:43,411 STRUCTURES AND STANDARDS. 3405 02:45:43,411 --> 02:45:48,817 AND WE HAVE CHEMICAL FORMATIONS 3406 02:45:48,817 --> 02:45:50,085 AND VARIANTS. 3407 02:45:50,085 --> 02:45:53,822 SO FOR THIS PURPOSE WE USED 3408 02:45:53,822 --> 02:46:00,829 CHEMICAL REACTIONS STORED IN THE 3409 02:46:00,829 --> 02:46:11,272 DATABASE AND THIS MAP HAS 3410 02:46:21,716 --> 02:46:32,193 VARIABLES AND WE LOOKED AT THE 3411 02:46:40,635 --> 02:46:46,007 DATA AND THEN WE DID VALIDATE 3412 02:46:46,007 --> 02:46:52,614 THIS DISCOVERY OURSELVES BUT WE 3413 02:46:52,614 --> 02:46:59,954 FOUND THE NEW REACTION WHICH IS 3414 02:46:59,954 --> 02:47:04,826 LISTED AND THEN WE EXTENDED THIS 3415 02:47:04,826 --> 02:47:12,434 APPROACH ON THE SPACE BUILT ON 3416 02:47:12,434 --> 02:47:22,310 THE VARIABLES AND WE USED THE 3417 02:47:22,310 --> 02:47:28,183 DATA AND WITH THE MODEL 3418 02:47:28,183 --> 02:47:36,558 DISPLAYED WE USED THE OUTPUT 3419 02:47:36,558 --> 02:47:38,993 PAST THE CHEMICAL STRUCTURE SO 3420 02:47:38,993 --> 02:47:44,833 THE TECHNOLOGY CAN BE USED. 3421 02:47:44,833 --> 02:47:52,607 SO, NOW I'D LIKE TO SPEAK ABOUT 3422 02:47:52,607 --> 02:47:56,811 OTHER MOLECULES. 3423 02:47:56,811 --> 02:48:02,217 THE PROBLEM WITH CONVENTIONAL 3424 02:48:02,217 --> 02:48:02,851 PIPELINE WITH DIFFERENT 3425 02:48:02,851 --> 02:48:08,022 POPULATION IS IT'S TIME 3426 02:48:08,022 --> 02:48:09,290 CONSUMING. 3427 02:48:09,290 --> 02:48:11,059 CERTAINLY IT'S A FUNCTION OF THE 3428 02:48:11,059 --> 02:48:21,603 MOLECULE INTRODUCED BUT WE HAD 4 3429 02:48:21,836 --> 02:48:30,111 MILLION COMPOUNDS AND WE DECIDED 3430 02:48:30,111 --> 02:48:33,715 TO BUILD A NEURAL NETWORK WHICH 3431 02:48:33,715 --> 02:48:40,421 SKIPS THE ENUMERATION STEP AND 3432 02:48:40,421 --> 02:48:40,722 PREPARATION. 3433 02:48:40,722 --> 02:48:45,593 SO I WOULD LIKE TO GIVE DETAILS 3434 02:48:45,593 --> 02:48:52,600 OF THIS NEURAL NETWORK GIVEN IN 3435 02:48:52,600 --> 02:49:02,777 THE REFERENCE IN THE PAPER. 3436 02:49:02,777 --> 02:49:04,813 THIS IS AN ILLUSTRATION AND THIS 3437 02:49:04,813 --> 02:49:15,089 MODEL IS ON THE SMALL SET WITH 3438 02:49:15,089 --> 02:49:17,358 90 MILLION DELS. 3439 02:49:17,358 --> 02:49:22,597 YOU MAY SEE THE MAPS BUILT WITH 3440 02:49:22,597 --> 02:49:26,301 THE NEURAL NETWORK AND WHAT IS 3441 02:49:26,301 --> 02:49:32,841 IDENTICAL TO ME MAP BUILT FOR 3442 02:49:32,841 --> 02:49:36,945 ENUMERATED CHEMICAL STRUCTURES. 3443 02:49:36,945 --> 02:49:47,488 FINALLY, RECENTLY WE PERFORMED 3444 02:49:50,892 --> 02:49:55,864 A BENCHMARK WITH CONSIDERATIONS 3445 02:49:55,864 --> 02:49:57,098 COMPARED THE INITIAL SPACE IN 3446 02:49:57,098 --> 02:50:00,401 THE SPACE IN THE MAP. 3447 02:50:00,401 --> 02:50:10,144 WE USED MORE THAN 100 SETS FROM 3448 02:50:10,144 --> 02:50:20,088 CHEMBL AND THIS REDUCED THE 3449 02:50:20,088 --> 02:50:23,558 BENCHMARK STUDY AND FOR RAY 3450 02:50:23,558 --> 02:50:28,263 GIVEN DATA SET, ALL TECHNIQUES 3451 02:50:28,263 --> 02:50:36,437 ARE PARTICULARLY AT THE SAME 3452 02:50:36,437 --> 02:50:40,275 LEVEL AND NEVER HAVE 100% 3453 02:50:40,275 --> 02:50:43,144 PRESERVATION IT'S 50% OR 70%. 3454 02:50:43,144 --> 02:50:49,918 BUT GTM OUT PERFORMS OTHER 3455 02:50:49,918 --> 02:50:52,620 METHODS WHEN WE PROJECT THE 3456 02:50:52,620 --> 02:50:52,820 FORM. 3457 02:50:52,820 --> 02:50:56,524 THAT'S WHY OUR CHOICE OF GTM IS 3458 02:50:56,524 --> 02:50:57,325 WELL FUNDED. 3459 02:50:57,325 --> 02:51:00,828 SO VERY LAST SLIDE. 3460 02:51:00,828 --> 02:51:04,632 I WOULD LIKE JUST TO SUMMARIZE 3461 02:51:04,632 --> 02:51:11,406 THE GTM AND HAS DATA PROBABILITY 3462 02:51:11,406 --> 02:51:12,340 DISTRIBUTION SUITABLE OR DATA 3463 02:51:12,340 --> 02:51:14,442 ANALYSIS AND WE HAVE DATA POINTS 3464 02:51:14,442 --> 02:51:17,812 TO GO FROM INDIVIDUAL 3465 02:51:17,812 --> 02:51:24,819 PROJECTIONS TO DAILY 3466 02:51:24,819 --> 02:51:25,219 DISTRIBUTION. 3467 02:51:25,219 --> 02:51:28,623 THE MODEL IS REPRESENTATIVE OF 3468 02:51:28,623 --> 02:51:32,560 THE SUBSET. 3469 02:51:32,560 --> 02:51:36,030 THE LANDSCAPES CAN BE USED FOR 3470 02:51:36,030 --> 02:51:43,471 REGRESSION OR SPECIFICATION TEST 3471 02:51:43,471 --> 02:51:49,210 AND ENABLES CONVERGENCE OF 3472 02:51:49,210 --> 02:51:51,646 LIBRARIES AND FINALLY ET 3473 02:51:51,646 --> 02:51:54,315 FACILITATES DE NOVO DESIGN. 3474 02:51:54,315 --> 02:51:55,850 I'D LIKE TO THANK MY 3475 02:51:55,850 --> 02:52:01,122 COLLABORATORS WHO PARTICIPATED 3476 02:52:01,122 --> 02:52:04,192 IN THIS WORK AND OUR PARTNERS 3477 02:52:04,192 --> 02:52:08,830 AND WOULD LIKE TO THANK YOU FOR 3478 02:52:08,830 --> 02:52:11,399 YOUR ATTENTION. 3479 02:52:11,399 --> 02:52:15,737 >> 3480 02:52:15,737 --> 02:52:18,106 >> THANK YOU VERY MUCH FOR AN 3481 02:52:18,106 --> 02:52:18,773 EXCELLENT PRESENTATION. 3482 02:52:18,773 --> 02:52:21,609 AND ANALYSIS OF GTM AND WHAT IT 3483 02:52:21,609 --> 02:52:22,610 CAN DO. 3484 02:52:22,610 --> 02:52:25,680 THEY HAVE BEEN A FEW QUESTIONS 3485 02:52:25,680 --> 02:52:28,383 ON THE SIDEBAR WHICH YOU'LL READ 3486 02:52:28,383 --> 02:52:28,683 SOME TO YOU. 3487 02:52:28,683 --> 02:52:32,820 SOME YOU ANSWERED AS YOU 3488 02:52:32,820 --> 02:52:38,326 PROGRESSED. 3489 02:52:38,326 --> 02:52:41,496 ONE OF THE QUESTIONS THAT I'D 3490 02:52:41,496 --> 02:52:47,602 LIKE TO SELECT IS THE ZINC 3491 02:52:47,602 --> 02:52:49,170 DATABASE, WHY DID YOU USE CHEMBL 3492 02:52:49,170 --> 02:52:59,847 AS A REFERENCE LIBRARY FOR DEL. 3493 02:52:59,847 --> 02:53:06,487 >> WE CONSIDERED A TASK OF 3494 02:53:06,487 --> 02:53:07,789 COLLECTION OF DEL FOR PRIMARY 3495 02:53:07,789 --> 02:53:08,056 SCREENING. 3496 02:53:08,056 --> 02:53:11,592 WHEN THE TARGET IS NOT KNOWN SO 3497 02:53:11,592 --> 02:53:14,462 WE NEED TO HAVE AS A REFERENCE A 3498 02:53:14,462 --> 02:53:16,631 MORE DIVERSE DATA COLLECTION. 3499 02:53:16,631 --> 02:53:19,767 THAT'S WHY. 3500 02:53:19,767 --> 02:53:20,268 WE CONSIDERED THE DATA 3501 02:53:20,268 --> 02:53:22,603 COLLECTION. 3502 02:53:22,603 --> 02:53:26,474 >> THANK YOU. 3503 02:53:26,474 --> 02:53:29,544 ANY COMMENT ON WHETHER AND HOW 3504 02:53:29,544 --> 02:53:33,448 YOU CAN USE GTM TO ADDRESS THE 3505 02:53:33,448 --> 02:53:39,454 QUESTION OF THE CLIPS? 3506 02:53:39,454 --> 02:53:47,795 >> I WOULD SAY GTM WE HAVE WORK 3507 02:53:47,795 --> 02:53:50,731 ON THE TOPIC BUT GTM IS NOT THE 3508 02:53:50,731 --> 02:53:56,204 BEST TOOL TO IDENTIFY THE CASE 3509 02:53:56,204 --> 02:53:58,106 BECAUSE YOU CAN'T WORK WITH THE 3510 02:53:58,106 --> 02:54:03,277 DISTRIBUTION AND CAPTURES MOSTLY 3511 02:54:03,277 --> 02:54:03,644 THIS LANDSCAPE. 3512 02:54:03,644 --> 02:54:04,579 >> THANK YOU. 3513 02:54:04,579 --> 02:54:06,714 AND THE LAST QUESTION GIVEN THE 3514 02:54:06,714 --> 02:54:07,782 TIME CONSTRAINTS, WHEN YOU 3515 02:54:07,782 --> 02:54:15,356 COMPARE GTM WITH OTHER, THE 3516 02:54:15,356 --> 02:54:16,624 PRESENTATION IS 2-D BUT 3517 02:54:16,624 --> 02:54:24,832 COMPOUNDS ARE 3-D. 3518 02:54:24,832 --> 02:54:32,807 >> WE PUBLISHED PAPERED ON 3519 02:54:32,807 --> 02:54:43,484 CHEMICAL REACTIONS AND WE HAVE 3520 02:54:45,686 --> 02:54:48,723 PROTEINS FOR A SERIES OF 3521 02:54:48,723 --> 02:54:53,961 CHEMICAL REACTION AND CAN BE 3522 02:54:53,961 --> 02:54:58,633 USED FOR ANALYSIS OF COMPLEX 3-D 3523 02:54:58,633 --> 02:55:01,235 STRUCTURES. 3524 02:55:01,235 --> 02:55:01,903 >> OKAY. 3525 02:55:01,903 --> 02:55:02,970 THANK YOU. 3526 02:55:02,970 --> 02:55:05,673 IF YOU HAVE TIME, WOULD YOU 3527 02:55:05,673 --> 02:55:07,275 PLEASE ANSWER SOME QUESTIONS IN 3528 02:55:07,275 --> 02:55:11,345 THE Q&A ONLINE YOU COULD KINDLY 3529 02:55:11,345 --> 02:55:12,747 DO THAT AND THANK YOU FOR AN 3530 02:55:12,747 --> 02:55:15,750 EXCELLENT TALK. 3531 02:55:15,750 --> 02:55:19,086 LET'S MOVE TO THE NEXT SPEAKER 3532 02:55:19,086 --> 02:55:22,056 IN THE SESSION, GISBERT 3533 02:55:22,056 --> 02:55:22,323 SCHNEIDER. 3534 02:55:22,323 --> 02:55:27,595 HE IS ONE OF THE HIGHEST CITED 3535 02:55:27,595 --> 02:55:28,729 SPECIALISTS IN CHEMISTRY. 3536 02:55:28,729 --> 02:55:31,432 HE HAS SPEARHEADED THE FUSION OF 3537 02:55:31,432 --> 02:55:33,000 MACHINE LEARNING AND A.I. WITH 3538 02:55:33,000 --> 02:55:35,236 MEDICINAL AND BIO ORGANIC 3539 02:55:35,236 --> 02:55:40,208 CHEMISTRY AND CREATED WITH 3540 02:55:40,208 --> 02:55:41,475 COINING THE TERMS AND NOT JUST 3541 02:55:41,475 --> 02:55:43,511 CREDITED WITH TERMS BUT HAS DONE 3542 02:55:43,511 --> 02:55:45,413 A LOT OF RESEARCH IN THESE AREAS 3543 02:55:45,413 --> 02:55:48,349 AND DEVELOPED GROUNDBREAKING 3544 02:55:48,349 --> 02:55:50,384 ALGORITHMS FOR MOLECULAR DESIGN 3545 02:55:50,384 --> 02:55:52,820 LEADING TO PROFOUND 3546 02:55:52,820 --> 02:55:56,324 TRANSFORMATION IN THE WAY WE 3547 02:55:56,324 --> 02:56:00,828 ATTEMPT TO DISCOVER NEW DRUGS. 3548 02:56:00,828 --> 02:56:05,700 HE'S A PREVENTION WITH 3549 02:56:05,700 --> 02:56:10,204 COMPUTERIZED DESIGN IN ETH 3550 02:56:10,204 --> 02:56:19,347 ZURICH AND HAS RECEIVED MULTIPLE 3551 02:56:19,347 --> 02:56:19,580 AWARDS. 3552 02:56:19,580 --> 02:56:22,617 HIS TALK IS ON DENOVEE DESIGN. 3553 02:56:22,617 --> 02:56:25,753 >> THANK YOU. 3554 02:56:25,753 --> 02:56:27,855 IT'S A PLEASURE DISCUSSING SOME 3555 02:56:27,855 --> 02:56:31,092 OF OUR WORK AND OUR IDEAS 3556 02:56:31,092 --> 02:56:36,063 REVOLVING AROUND DE NOVO DESIGN. 3557 02:56:36,063 --> 02:56:37,431 WE'RE ACTUALLY COMBINING MACHINE 3558 02:56:37,431 --> 02:56:40,034 LEARNING AND APPLIED MEDICINAL 3559 02:56:40,034 --> 02:56:42,903 CHEMISTRY UNDER ONE ROOF IN ONE 3560 02:56:42,903 --> 02:56:45,940 LAB AND MAYBE NEW MOLECULE AND 3561 02:56:45,940 --> 02:56:46,607 CHALLENGE A.I. AND MACHINE 3562 02:56:46,607 --> 02:56:48,142 LEARNING TOOLS UNTIL THEY BREAK 3563 02:56:48,142 --> 02:56:52,647 TO SEE WHICH DEGREE CAN WE USE A 3564 02:56:52,647 --> 02:56:53,281 CERTAIN TECHNOLOGY FOR DE NOVO 3565 02:56:53,281 --> 02:56:56,117 DESIGN. 3566 02:56:56,117 --> 02:57:00,154 NOW, WHY IS IT SO DIFFICULT TO 3567 02:57:00,154 --> 02:57:02,823 COME UP WITH NEW DRUGS? 3568 02:57:02,823 --> 02:57:04,825 ONE REASON IS THE SEARCH SPACE 3569 02:57:04,825 --> 02:57:07,828 IS VIRTUALLY INFINITE. 3570 02:57:07,828 --> 02:57:09,730 THE TASK IS TO PICK ONE MOLECULE 3571 02:57:09,730 --> 02:57:12,867 OR 10 MOLECULES, MAYBE 100 3572 02:57:12,867 --> 02:57:16,671 MOLECULES FROM A VIRTUALLY 3573 02:57:16,671 --> 02:57:20,474 INFINITE SEARCH SPACE. 3574 02:57:20,474 --> 02:57:21,909 TRADITIONALLY THIS HAS BEEN DONE 3575 02:57:21,909 --> 02:57:24,812 IT'S ONE OF THE DRIVING FORCES 3576 02:57:24,812 --> 02:57:27,615 OF DRUG DISCOVERIES AND VIRTUAL 3577 02:57:27,615 --> 02:57:29,483 SCREENING AS WE ALREADY HAVE 3578 02:57:29,483 --> 02:57:35,323 HEARD IN THE PRECEDING TALKS, 3579 02:57:35,323 --> 02:57:36,490 CAN DELIVER USEFUL COMPOUNDS WE 3580 02:57:36,490 --> 02:57:40,261 CAN TEST IN THE LAB. 3581 02:57:40,261 --> 02:57:41,896 NOW, DE NOVO DESIGN AND IN 3582 02:57:41,896 --> 02:57:44,065 PARTICULAR A PARTICULAR ASPECT 3583 02:57:44,065 --> 02:57:45,633 OF GENERATIVE A.I. WHICH I WILL 3584 02:57:45,633 --> 02:57:50,871 HIGHLIGHT IN MY TALK INVERTS 3585 02:57:50,871 --> 02:57:51,205 THIS CONCEPT. 3586 02:57:51,205 --> 02:57:52,640 WE'RE NO LONGER SCREENING FOR 3587 02:57:52,640 --> 02:57:56,477 NEW MOLECULES BUT CREATING AND 3588 02:57:56,477 --> 02:57:57,178 GENERATING THE MOLECULE WE WANT 3589 02:57:57,178 --> 02:57:57,445 TO SEE. 3590 02:57:57,445 --> 02:58:06,354 IN A WAY, WE CAN AVOID THE VAST 3591 02:58:06,354 --> 02:58:06,987 SEARCH SPACE BY ASKING A 3592 02:58:06,987 --> 02:58:09,490 COMPUTER ALGORITHM TO COME UP 3593 02:58:09,490 --> 02:58:12,493 WITH HOPEFULLY OPTIMAL OR AT 3594 02:58:12,493 --> 02:58:14,829 LEAST LOCALLY OPTIMAL SOLUTIONS 3595 02:58:14,829 --> 02:58:22,303 IN THIS MOLECULAR SPACE. 3596 02:58:22,303 --> 02:58:23,804 SO DE NOVO DESIGN RESHAPES THE 3597 02:58:23,804 --> 02:58:26,040 DRUG DISCOVERY INVERTING 3598 02:58:26,040 --> 02:58:26,941 SCREENING TO DESIGNING 3599 02:58:26,941 --> 02:58:28,676 MOLECULES. 3600 02:58:28,676 --> 02:58:31,112 IT HAS THE POTENTIAL TO 3601 02:58:31,112 --> 02:58:32,813 COMPLEMENT AND POTENTIALLY 3602 02:58:32,813 --> 02:58:33,714 REPLACE HIGH THROUGHPUT 3603 02:58:33,714 --> 02:58:34,281 SCREENING AT LEAST TO SOME 3604 02:58:34,281 --> 02:58:36,817 DEGREE. 3605 02:58:36,817 --> 02:58:40,654 THE MOST IMPORTANT ASPECT OF DE 3606 02:58:40,654 --> 02:58:43,023 NOVO DESIGN I THINK IS THE TOOLS 3607 02:58:43,023 --> 02:58:45,860 COME UP WITH SURPRISING NOVEL 3608 02:58:45,860 --> 02:58:48,362 IDEAS WHICH WE HAVEN'T 3609 02:58:48,362 --> 02:58:51,599 CONSIDERED OR A CHEMIST HASN'T 3610 02:58:51,599 --> 02:58:53,467 CONSIDERED AT THE TIME. 3611 02:58:53,467 --> 02:58:57,171 THE AHA EFFECT OF ALGORITHMS IN 3612 02:58:57,171 --> 02:58:58,672 MY VIEW IS ONE OF THE MOST 3613 02:58:58,672 --> 02:58:59,607 IMPORTANT ASSETS WE HAVE WHEN WE 3614 02:58:59,607 --> 02:59:02,510 USE THESE TOOLS. 3615 02:59:02,510 --> 02:59:06,414 BEFORE I DIVE INTO THE WORLD OF 3616 02:59:06,414 --> 02:59:07,047 DE NOVO DESIGN ALGORITHMS, I'D 3617 02:59:07,047 --> 02:59:08,616 LIKE TO MENTION THAT WE'RE 3618 02:59:08,616 --> 02:59:11,252 WORKING AT THE EDGE OF CHAOS 3619 02:59:11,252 --> 02:59:16,657 WHEN WE INTERFERE WITH LIVING 3620 02:59:16,657 --> 02:59:18,426 BIOLOGICAL ALGORITHMS THE 3621 02:59:18,426 --> 02:59:22,062 PROBLEMS OF DATA AND MODELS AND 3622 02:59:22,062 --> 02:59:28,035 THE PROBLEM OF NON LINEARITY AND 3623 02:59:28,035 --> 02:59:30,271 WE HEARD ABOUT ACTIVITY CLIFFS 3624 02:59:30,271 --> 02:59:32,440 IN THE PREVIOUS TALK AND THIS 3625 02:59:32,440 --> 02:59:33,941 BOILS DOWN TO PARTIAL 3626 02:59:33,941 --> 02:59:34,675 PREDICTABILITY. 3627 02:59:34,675 --> 02:59:38,813 WE HAVE TO CONCEDE OUR 3628 02:59:38,813 --> 02:59:41,015 IMPERFECT, INCOMPLETE 3629 02:59:41,015 --> 02:59:43,517 UNDERSTANDING OF THE MOLECULAR 3630 02:59:43,517 --> 02:59:47,555 PATHOLOGY AND MOLECULAR COURSES 3631 02:59:47,555 --> 02:59:48,589 OF DISEASE AND ACCEPTING THE 3632 02:59:48,589 --> 02:59:54,695 FACT WE WILL NOT BE ABLE TO MAKE 3633 02:59:54,695 --> 03:00:00,334 PERFECT PREDICTIONS BUT WE CAN 3634 03:00:00,334 --> 03:00:01,569 OF COURSE USE DE NOVO DESIGN 3635 03:00:01,569 --> 03:00:02,536 WITH A.I. IN A VERY FRUITFUL 3636 03:00:02,536 --> 03:00:08,642 MANNER. 3637 03:00:08,642 --> 03:00:09,743 AND BECAUSE DRUG DISCOVERY IS 3638 03:00:09,743 --> 03:00:17,852 NOT A CLASSIC ENGINEERING 3639 03:00:17,852 --> 03:00:18,552 DISCIP 3640 03:00:18,552 --> 03:00:20,821 DISCIPLINE WE HAVE TO GO THROUGH 3641 03:00:20,821 --> 03:00:22,556 THE DESIGN CYCLE HIGHLIGHTED 3642 03:00:22,556 --> 03:00:26,560 HERE IN THIS CARTOON WE START 3643 03:00:26,560 --> 03:00:29,330 WITH HYPOTHESIS AND IN CHEMISTRY 3644 03:00:29,330 --> 03:00:32,833 IT'S A MOLECULAR STRUCTURE WE 3645 03:00:32,833 --> 03:00:33,567 THEN SYNTHESIZE IN THE LAB AND 3646 03:00:33,567 --> 03:00:38,606 THE MOLECULES HAVE TO BE TESTED 3647 03:00:38,606 --> 03:00:39,473 AND VALIDATED BASED ON THE 3648 03:00:39,473 --> 03:00:42,176 RESULTS WE CAN RELEARN AND 3649 03:00:42,176 --> 03:00:44,144 REFINE OUR HYPOTHESIS. 3650 03:00:44,144 --> 03:00:46,380 OVER THE PAST YEARS WE'VE SEEN A 3651 03:00:46,380 --> 03:00:49,884 CLEAR MOVE FROM HUMAN GENERATED 3652 03:00:49,884 --> 03:00:53,120 HYPOTHESIS TO MACHINE GENERATED 3653 03:00:53,120 --> 03:00:55,289 HYPOTHESES AND LEARNING. 3654 03:00:55,289 --> 03:00:58,325 AT THE SAME TIME ROBOTS TAKE 3655 03:00:58,325 --> 03:00:59,560 OVER CERTAIN PARTS IN THE 3656 03:00:59,560 --> 03:01:00,094 LABORATORY. 3657 03:01:00,094 --> 03:01:02,062 THE IDEA AND CONCEPT IS ALREADY 3658 03:01:02,062 --> 03:01:05,933 OUT THERE TO HAVE FULLY 3659 03:01:05,933 --> 03:01:07,801 AUTOMATED THIS DISCOVERY LAB AND 3660 03:01:07,801 --> 03:01:08,469 WE'LL HEAR ABOUT SOME OF THESE 3661 03:01:08,469 --> 03:01:15,276 TODAY AND TOMORROW. 3662 03:01:15,276 --> 03:01:16,744 CHEMISTRY AND MEDICINAL 3663 03:01:16,744 --> 03:01:18,546 CHEMISTRY ARE LEFT BEHIND WHEN 3664 03:01:18,546 --> 03:01:23,517 IT COMES TO AUTOMATED DESIGN. 3665 03:01:23,517 --> 03:01:24,818 HERE'S AN EXAMPLE. 3666 03:01:24,818 --> 03:01:30,357 ON THEN LEFT SIDE YOU SEE A 3667 03:01:30,357 --> 03:01:33,060 GENERATIVE MODEL OF SUCH A 3668 03:01:33,060 --> 03:01:35,162 HALLWAY WITH COLUMNS AND ON THE 3669 03:01:35,162 --> 03:01:40,534 RIGHT A CONCRETE PRINTER WHICH 3670 03:01:40,534 --> 03:01:44,705 HAS BEEN BUILT BY MY COLLEAGUE 3671 03:01:44,705 --> 03:01:47,174 AND ONCE YOU HAVE THE DESIGN THE 3672 03:01:47,174 --> 03:01:49,109 3-D PRINTERS IMMEDIATELY 3673 03:01:49,109 --> 03:01:51,478 GENERATE THE BUILDING BLOCKS FOR 3674 03:01:51,478 --> 03:01:56,684 NEW ARCHITECTURAL DESIGNS. 3675 03:01:56,684 --> 03:02:00,688 IF IT WERE ONLY SO EASY WITH 3676 03:02:00,688 --> 03:02:01,288 CHEMICAL SYNTHESIS I'D SHOW 3677 03:02:01,288 --> 03:02:02,957 EXAMPLES OF WHAT WORKS BE AND 3678 03:02:02,957 --> 03:02:04,592 DOESN'T IN CHEMISTRY AND HOW 3679 03:02:04,592 --> 03:02:06,860 A.I. COULD POTENTIALLY OVERCOME 3680 03:02:06,860 --> 03:02:08,963 THIS BOTTLENECK OF CHEMICAL 3681 03:02:08,963 --> 03:02:09,530 SYNTHESIS IN THE DISCOVERY 3682 03:02:09,530 --> 03:02:15,302 CYCLE. 3683 03:02:15,302 --> 03:02:17,338 FOR DE NOVO DESIGN WE HAVE TO 3684 03:02:17,338 --> 03:02:19,373 HAVE STRUCTURAL GENERATION AND 3685 03:02:19,373 --> 03:02:21,508 SCORING AND SELECTION AND 3686 03:02:21,508 --> 03:02:24,545 OPTIMIZATION OF THE HYPOTHESIS. 3687 03:02:24,545 --> 03:02:26,981 THERE'S VARIOUS TOOLS AVAILABLE 3688 03:02:26,981 --> 03:02:30,551 TO GENERATE NEW STRUCTURES. 3689 03:02:30,551 --> 03:02:31,418 SCORING AND SELECTION IS THE 3690 03:02:31,418 --> 03:02:35,489 MOST CRITICAL AND HERE WE HAVE 3691 03:02:35,489 --> 03:02:38,058 THE GREATEST NEED FOR NOVEL 3692 03:02:38,058 --> 03:02:41,028 IDEAS AND PROSPECTIVE RESEARCH 3693 03:02:41,028 --> 03:02:42,763 AND FINALLY IN PARTICULAR 3694 03:02:42,763 --> 03:02:44,431 CERTAIN ACTIVE LEARNING 3695 03:02:44,431 --> 03:02:44,965 ALGORITHMS AVAILABLE FOR 3696 03:02:44,965 --> 03:02:48,802 OPTIMIZATION. 3697 03:02:48,802 --> 03:02:52,506 LET'S START WITH METHOD 1. 3698 03:02:52,506 --> 03:02:54,274 I'LL PRESENT FOUR DE NOVO DESIGN 3699 03:02:54,274 --> 03:02:56,810 METHODS AND I'LL START WITH A 3700 03:02:56,810 --> 03:02:57,711 HISTORIC ONE. 3701 03:02:57,711 --> 03:03:03,050 BECAUSE IT'S STILL QUIT ACTUAL 3702 03:03:03,050 --> 03:03:05,152 AND QUITE USEFUL NAMELY, DE NOVO 3703 03:03:05,152 --> 03:03:08,288 DESIGN IN A LOW DATA SCENARIO 3704 03:03:08,288 --> 03:03:10,057 WHEN THERE'S MAYBE ONE COMPOUND 3705 03:03:10,057 --> 03:03:13,761 AVAILABLE YOU WISH TO MIMIC AND 3706 03:03:13,761 --> 03:03:16,397 MAKE A SCAFFOLD OF. 3707 03:03:16,397 --> 03:03:18,132 THE IDEA IS IT MAKE SMALL 3708 03:03:18,132 --> 03:03:20,834 MOLECULAR BUILDING BLOCKS AND 3709 03:03:20,834 --> 03:03:23,637 ANNOTATE THEM IN THE DATABASE TO 3710 03:03:23,637 --> 03:03:24,805 WHICH POTENTIAL CHEMICAL 3711 03:03:24,805 --> 03:03:25,572 TRANSFORMATION THEY COULD BE 3712 03:03:25,572 --> 03:03:26,340 USED IN. 3713 03:03:26,340 --> 03:03:28,942 A RULE-BASED ARCHITECTURE UNDER 3714 03:03:28,942 --> 03:03:33,280 LIES IS THE BASIS HERE OF THIS 3715 03:03:33,280 --> 03:03:34,415 MACHINE LEARNING TOOL. 3716 03:03:34,415 --> 03:03:36,817 YOU DEFINE BUILDING BLOCKS AND 3717 03:03:36,817 --> 03:03:39,720 CHEMICAL TRANSFORMATION AND CAN 3718 03:03:39,720 --> 03:03:41,221 ENUMERATE AND REITERATE THROUGH 3719 03:03:41,221 --> 03:03:44,892 THE CHEMICAL SPACES. 3720 03:03:44,892 --> 03:03:48,295 THESE APPROACHES ARE RULE-BASED 3721 03:03:48,295 --> 03:03:49,697 MOLECULAR DESIGN ARE EXTREMELY 3722 03:03:49,697 --> 03:03:50,898 VALUABLE WHEN ONLY LIMITED DATA 3723 03:03:50,898 --> 03:03:54,902 IS AVAILABLE AND AT THE SAME 3724 03:03:54,902 --> 03:03:56,804 TIME WITH A NEW DESIGN BECAUSE 3725 03:03:56,804 --> 03:03:59,406 WE USE BUILDING BLOCKS AND 3726 03:03:59,406 --> 03:04:03,110 CHEMICAL TRANSFORM S FOR CON 3727 03:04:03,110 --> 03:04:03,610 STRU 3728 03:04:03,610 --> 03:04:04,712 INSTRUCTING MOLECULES THEY COME 3729 03:04:04,712 --> 03:04:07,314 UM WITH A PATHWAY FOR EACH 3730 03:04:07,314 --> 03:04:11,318 MOLECULE TO BE FED DIRECTLY INTO 3731 03:04:11,318 --> 03:04:12,419 AN AUTOMATED SYNTHESIZER TO 3732 03:04:12,419 --> 03:04:21,428 PRODUCE THE MOLECULES GENERATED 3733 03:04:21,428 --> 03:04:23,464 WITH THE MACHINE LEARNING 3734 03:04:23,464 --> 03:04:23,731 APPROACH. 3735 03:04:23,731 --> 03:04:26,233 THE QUESTION AROSE THE 3736 03:04:26,233 --> 03:04:28,035 DIFFERENCE BETWEEN SCAFFOLD 3737 03:04:28,035 --> 03:04:28,736 TOPPING AND DE NOVO DESIGN 3738 03:04:28,736 --> 03:04:32,840 DESIGN AND YOU SEE THE FIRST DE 3739 03:04:32,840 --> 03:04:36,810 NOVO DESIGN DESIGN WITH A 3740 03:04:36,810 --> 03:04:42,249 MACHINE LEARNING DE NOVO 3741 03:04:42,249 --> 03:04:44,418 STRUCTURE AND AUTOMIZATION FROM 3742 03:04:44,418 --> 03:04:46,653 THE TEMPLATE MOLECULE AS 3743 03:04:46,653 --> 03:04:50,357 REFERENCE WHICH YOU WISH TO 3744 03:04:50,357 --> 03:04:51,024 MIMIC THE DE NOVO DESIGN TOOL 3745 03:04:51,024 --> 03:04:52,860 COMES UP WITH A SUGGESTION THAT 3746 03:04:52,860 --> 03:04:56,830 NEEDS TO BE SYNTHESIZES IN THE 3747 03:04:56,830 --> 03:04:58,198 LAB AND TESTED. 3748 03:04:58,198 --> 03:05:01,535 AND THIS IS FED BACK TO THE 3749 03:05:01,535 --> 03:05:04,805 LEARNING ALGORITHM FOR ITERATIVE 3750 03:05:04,805 --> 03:05:07,407 COMBINATORIAL OPTIMIZATION. 3751 03:05:07,407 --> 03:05:10,410 WHEN WE APPLY DE NOVO DESIGN 3752 03:05:10,410 --> 03:05:14,548 SUCH A FASHION IS AN INSTANCE OF 3753 03:05:14,548 --> 03:05:15,816 SCAFFOLD POPPING. 3754 03:05:15,816 --> 03:05:18,952 HERE'S ANOTHER EXAMPLE ONE CAN 3755 03:05:18,952 --> 03:05:22,689 USE DE NOVO DESIGN THE 3756 03:05:22,689 --> 03:05:24,124 RULE-BASED DE NOVO DESIGN TOOLS. 3757 03:05:24,124 --> 03:05:28,295 FOR EXAMPLE WHEN YOU WISH TO 3758 03:05:28,295 --> 03:05:34,802 LOOK AT A COMPLEX SCAFFOLD. 3759 03:05:34,802 --> 03:05:37,938 YOU SEE WHAT HAS BEEN 3760 03:05:37,938 --> 03:05:45,546 DE-ORPHANED AS A MENTHOL CHANNEL 3761 03:05:45,546 --> 03:05:47,981 BETA-BLOCKER BUT REQUIRES A 3762 03:05:47,981 --> 03:05:51,051 SYNTHESIS OF 11 LINEAR STEPS. 3763 03:05:51,051 --> 03:05:53,887 NOW, DE NOVO DESIGN HAS COME UP 3764 03:05:53,887 --> 03:06:00,794 WITH TWO NOVEL CHEMICAL 3765 03:06:00,794 --> 03:06:01,094 STRUCTURES. 3766 03:06:01,094 --> 03:06:02,629 THESE COULD ALL BE SYNTHESIZES 3767 03:06:02,629 --> 03:06:04,798 IN THREE STEPS AND HAVE DESIGN 3768 03:06:04,798 --> 03:06:08,168 ACTIVITY AS THE ORIGINAL 3769 03:06:08,168 --> 03:06:08,435 TEMPLATE. 3770 03:06:08,435 --> 03:06:09,636 AGAIN, DE NOVO DESIGN CAN HELP 3771 03:06:09,636 --> 03:06:16,610 US GENERATE NEW STRUCTURES THESE 3772 03:06:16,610 --> 03:06:17,945 ARE IP-FROM AN AND SYNTHESIZED 3773 03:06:17,945 --> 03:06:22,216 AND MIMIC IN THIS CASE COMPLEX 3774 03:06:22,216 --> 03:06:23,884 NATURAL PRODUCT TEMPLATE. 3775 03:06:23,884 --> 03:06:28,622 HERE'S A MORE RECENT EXAMPLE OF 3776 03:06:28,622 --> 03:06:32,025 THE APPLICATION OF RULE-BASED DE 3777 03:06:32,025 --> 03:06:36,830 NOVO DESIGN AND OTHER PRODUCT 3778 03:06:36,830 --> 03:06:42,135 WAS CONVERTED INTO THIS DE NOVO 3779 03:06:42,135 --> 03:06:49,376 STRUCTURE AND IT'S THE MOST COX 3780 03:06:49,376 --> 03:06:59,920 1 INHIBITOR TO DATE AND WE SE 3781 03:07:05,926 --> 03:07:13,300 SEE'S -- IT SHOWCASES THE TOOLS 3782 03:07:13,300 --> 03:07:19,239 OF THE STRUCTURES KNOWN IN THE 3783 03:07:19,239 --> 03:07:20,807 DOMAIN BUT WOULD HAVING SEEN THE 3784 03:07:20,807 --> 03:07:22,576 STRUCTURES AND LOOKING WHY THE 3785 03:07:22,576 --> 03:07:23,644 DE NOVO DESIGN DOESN'T ACT LIKE 3786 03:07:23,644 --> 03:07:32,219 IT BUT IS A SELECTIVE COX 1 3787 03:07:32,219 --> 03:07:33,153 INHIBITOR OF COURSE THEN 3788 03:07:33,153 --> 03:07:34,221 EXPERIMENTAL STRUCTURE 3789 03:07:34,221 --> 03:07:39,493 DETERMINATION IS OF THE ESSENCE 3790 03:07:39,493 --> 03:07:45,766 TO HAVE HARD FACTS TO LEAD 3791 03:07:45,766 --> 03:07:46,833 OPTIMIZA 3792 03:07:46,833 --> 03:07:47,501 OPTIMIZATION. 3793 03:07:47,501 --> 03:07:51,371 HERE'S A LINK BETWEEN DESIGN AND 3794 03:07:51,371 --> 03:07:51,638 FOLLOW-UP. 3795 03:07:51,638 --> 03:07:55,776 METHOD TWO, WE'RE MOVING ON IN 3796 03:07:55,776 --> 03:07:58,812 HISTORY MORE RECENTLY CHEMICAL 3797 03:07:58,812 --> 03:08:00,147 LANGUAGE MODELS HAVE BEEN 3798 03:08:00,147 --> 03:08:00,547 FASHIONABLE. 3799 03:08:00,547 --> 03:08:04,818 WE HAD A CLOSE LOOK AT CHEMICAL 3800 03:08:04,818 --> 03:08:06,954 LANGUAGE MODELS TO LEARN THE 3801 03:08:06,954 --> 03:08:12,159 SYNTAX OF DRUG LIKE MOLECULE 3802 03:08:12,159 --> 03:08:14,394 FROM THE REPRESENTATION. 3803 03:08:14,394 --> 03:08:16,363 THE FIRST STEP OF CHEMICAL 3804 03:08:16,363 --> 03:08:20,434 LANGUAGE MODELS IS TO DEFINE A 3805 03:08:20,434 --> 03:08:22,469 TRAINING DATA SET OF DATA 3806 03:08:22,469 --> 03:08:23,737 SUITABLE FOR THE PARTICULAR TASK 3807 03:08:23,737 --> 03:08:27,874 REPRESENT THESE MOLECULES AS IN 3808 03:08:27,874 --> 03:08:32,813 THIS CASE STREAMS AND THEN IN AN 3809 03:08:32,813 --> 03:08:35,582 AUTO REGRESSIVE FASHION HAVE THE 3810 03:08:35,582 --> 03:08:38,118 CLM TOOL UNDERSTAND, IF I MAY 3811 03:08:38,118 --> 03:08:40,821 SAY SO, THE SYNTAX OF SMALL 3812 03:08:40,821 --> 03:08:43,924 STREAMS TO REPRODUCE THE INPUT 3813 03:08:43,924 --> 03:08:44,825 MOLECULES OF THE TRAINING DATA. 3814 03:08:44,825 --> 03:08:47,794 ONCE DONE, WE CAN USE SUCH A 3815 03:08:47,794 --> 03:08:51,164 TOOL TO EMIT TO GENERATE NEW 3816 03:08:51,164 --> 03:08:53,000 MOLECULES BASED ON THE INTERNAL 3817 03:08:53,000 --> 03:08:56,803 PROBABILITY DISTRIBUTIONS 3818 03:08:56,803 --> 03:08:57,738 LEARNED. 3819 03:08:57,738 --> 03:09:00,707 NOW ALL THESE TOOLS, THESE DEEP 3820 03:09:00,707 --> 03:09:01,808 LEARNING TOOLS REQUIRE DATA TO 3821 03:09:01,808 --> 03:09:02,976 LEARN FROM. 3822 03:09:02,976 --> 03:09:05,912 WE TYPICALLY RELY ON CHEMICAL 3823 03:09:05,912 --> 03:09:08,048 DATABASES, SUPPLY CATALOGS, 3824 03:09:08,048 --> 03:09:16,256 PATTERNS OR AS WE HEARD FROM 3825 03:09:16,256 --> 03:09:17,657 ALEXANDRE VARNEK COMPOUND 3826 03:09:17,657 --> 03:09:18,825 SAMPLES. 3827 03:09:18,825 --> 03:09:20,727 COMPARED TO FACE RECOGNITION, 3828 03:09:20,727 --> 03:09:24,431 COMPUTATIONAL DRUG DESIGN, 3829 03:09:24,431 --> 03:09:26,466 SURFACE FROM DATA THEREFORE IT'S 3830 03:09:26,466 --> 03:09:32,806 ESSENTIAL WE DEFINE CONTEXT 3831 03:09:32,806 --> 03:09:36,410 SPECIFIC DATA SELECTIONS FOR 3832 03:09:36,410 --> 03:09:39,646 WHAT IS AT HAND AND REPRESENT 3833 03:09:39,646 --> 03:09:42,115 THE MOLECULES. 3834 03:09:42,115 --> 03:09:42,949 HUMAN INGENUITY AND KNOWLEDGE 3835 03:09:42,949 --> 03:09:44,818 PLAYS AN ESSENTIAL PART. 3836 03:09:44,818 --> 03:09:46,820 WE CANNOT EXPECT A DEEP LEARNING 3837 03:09:46,820 --> 03:09:48,555 TOOL FOR EXAMPLE A CHEMICAL 3838 03:09:48,555 --> 03:09:49,856 LANGUAGE MODEL TO LEARN 3839 03:09:49,856 --> 03:09:52,826 EVERYTHING THERE IS ABOUT 3840 03:09:52,826 --> 03:09:55,262 CHEMISTRY AND ABOUT THE RELEVANT 3841 03:09:55,262 --> 03:09:56,630 MOLECULAR PROPERTIES BY FEEDING 3842 03:09:56,630 --> 03:10:00,534 IN MORE AND MORE DATA. 3843 03:10:00,534 --> 03:10:02,903 THIS IS AN OBSERVATION OVER THE 3844 03:10:02,903 --> 03:10:04,805 PAST YEARS WORKING ONE THESE 3845 03:10:04,805 --> 03:10:05,439 TOOLS. 3846 03:10:05,439 --> 03:10:07,074 A NATURAL LANGUAGE TO LEARN FOR 3847 03:10:07,074 --> 03:10:10,811 SUCH A CHEMICAL LANGUAGE MODEL 3848 03:10:10,811 --> 03:10:14,081 IS THE AMINO ACID CODE. 3849 03:10:14,081 --> 03:10:19,419 HERE'S AN EXAMPLE OF AN 3850 03:10:19,419 --> 03:10:21,154 AUTOMATED DESIGN TEST PIPELINE 3851 03:10:21,154 --> 03:10:23,924 WHERE WE TRAINED CHEMICAL 3852 03:10:23,924 --> 03:10:26,927 LANGUAGE MODEL ON SEQUENCES. 3853 03:10:26,927 --> 03:10:28,829 WE WANTED TO GENERATE AND ASK 3854 03:10:28,829 --> 03:10:32,799 THE COMPUTER TO GENERATE NEW 3855 03:10:32,799 --> 03:10:34,835 ANTI-CANCER PEPTIDES. 3856 03:10:34,835 --> 03:10:39,439 THE SYSTEM WAS TRAINED WITH 3857 03:10:39,439 --> 03:10:46,813 KNOWN HELICAL SEQUENCES AND 3858 03:10:46,813 --> 03:10:48,815 PEPTIDES. 3859 03:10:48,815 --> 03:10:52,319 AND CANCER CELLS IT'S WHAT YOU 3860 03:10:52,319 --> 03:10:54,187 SEE HERE IS BREAST CANCER CELL 3861 03:10:54,187 --> 03:11:04,397 AND MCF7 CELL THAT CONTAINS 3862 03:11:04,397 --> 03:11:06,433 FLUORESCENT DYE AND WHEN THE 3863 03:11:06,433 --> 03:11:07,701 PEPTIDE IS APPLIED THE CELL 3864 03:11:07,701 --> 03:11:14,141 LEAKS AND DIES. 3865 03:11:14,141 --> 03:11:16,810 AND IT TURNED OUT TO BE A CANCER 3866 03:11:16,810 --> 03:11:20,046 CELL SELECTIVE PEPTIDE THAT 3867 03:11:20,046 --> 03:11:24,751 LEAVES HEALTHY CELLS INTACT. 3868 03:11:24,751 --> 03:11:26,286 WHEN WE TALK ABOUT FULL 3869 03:11:26,286 --> 03:11:28,822 AUTOMATION AND INTEGRATING THE 3870 03:11:28,822 --> 03:11:34,861 DESIGN THEY TEST PARTS OF SUCH 3871 03:11:34,861 --> 03:11:40,433 AN AUTO TON WE CAN BUILD 3872 03:11:40,433 --> 03:11:45,572 MACHINES AND THIS IS ON A DESK 3873 03:11:45,572 --> 03:11:46,873 TOP AND PORTABLE DE NOVO DESIGN 3874 03:11:46,873 --> 03:11:49,509 LABORATORY GUIDED BY A CHEMICAL 3875 03:11:49,509 --> 03:11:52,812 LANGUAGE MODEL WHICH TRAINED ON 3876 03:11:52,812 --> 03:11:54,881 VENDOR CATALOGS. 3877 03:11:54,881 --> 03:11:59,085 ON SYNTHESIZABLE MOLECULES WE 3878 03:11:59,085 --> 03:12:00,387 WANTED IT TO LEARN ABOUT STOCK 3879 03:12:00,387 --> 03:12:02,589 MARKET ABILITY OF THE DESIGNS. 3880 03:12:02,589 --> 03:12:03,557 OVER ALL IN THIS PARTICULAR 3881 03:12:03,557 --> 03:12:08,195 APPLICATION WE HAD MORE THAN 50% 3882 03:12:08,195 --> 03:12:09,796 SUCCESSFUL SYNTHESISES AND MORE 3883 03:12:09,796 --> 03:12:14,901 THAN 50% OF THOSE MOLECULES 3884 03:12:14,901 --> 03:12:16,336 SYNTHESIZED HAD THE DESIRED 3885 03:12:16,336 --> 03:12:16,836 BIOLOGICAL ACTIVITY. 3886 03:12:16,836 --> 03:12:20,540 THAT'S A TYPICAL OUTCOME FOR DE 3887 03:12:20,540 --> 03:12:22,842 NOVO DESIGN WITH CHEMICAL 3888 03:12:22,842 --> 03:12:24,678 LANGUAGE MODELS AS WELL AS WITH 3889 03:12:24,678 --> 03:12:26,279 RULE BASED TOOLS. 3890 03:12:26,279 --> 03:12:28,815 WE CAN EXPECT -- ONE CAN EXPECT 3891 03:12:28,815 --> 03:12:32,819 AT LEAST A 50% IT'S A 3892 03:12:32,819 --> 03:12:34,721 CONSERVATIVE ESTIMATE CHANCE OF 3893 03:12:34,721 --> 03:12:35,288 BEING SYNTHESIZABLE AND BIO 3894 03:12:35,288 --> 03:12:40,794 ACTIVE. 3895 03:12:40,794 --> 03:12:43,530 WHEN WE LOOK AT THE MOLECULE 3896 03:12:43,530 --> 03:12:45,732 SUGGESTED BY THE MODEL WE SEE 3897 03:12:45,732 --> 03:12:47,400 APPROXIMATELY 50% OF THESE 3898 03:12:47,400 --> 03:12:48,802 SUBJECTED DESIGNS ARE ACTUALLY 3899 03:12:48,802 --> 03:12:51,304 NEW AND THE OTHERS ARE KNOWN IN 3900 03:12:51,304 --> 03:12:52,539 ONE OR THE OTHER WAY. 3901 03:12:52,539 --> 03:12:55,642 I THINK THIS IS A HEALTHY MIX 3902 03:12:55,642 --> 03:12:59,479 BECAUSE WE DON'T WANT TO CREATE 3903 03:12:59,479 --> 03:13:02,849 JUST ANY ARBITRARY NEW MOLECULE 3904 03:13:02,849 --> 03:13:05,252 BUT SEE THESE NEW MOLECULES AS 3905 03:13:05,252 --> 03:13:08,822 SIMILAR TO WHAT IS KNOWN BECAUSE 3906 03:13:08,822 --> 03:13:10,857 THIS INCREASES THE CHANCE THEY 3907 03:13:10,857 --> 03:13:16,830 ARE SYNTHESIZABLE AND HAVE THE 3908 03:13:16,830 --> 03:13:17,297 DESIRED BIO ACTIVITY. 3909 03:13:17,297 --> 03:13:21,835 METHOD THREE. 3910 03:13:21,835 --> 03:13:27,007 HERE I PRODUCED THE THIRD 3911 03:13:27,007 --> 03:13:30,644 DIMENSION BY GRAPH PROCESSING. 3912 03:13:30,644 --> 03:13:34,848 THE CHEMICAL WORLD IS NOT FLAT 3913 03:13:34,848 --> 03:13:36,650 AND WE NEED TO CONSIDER STEREO 3914 03:13:36,650 --> 03:13:39,452 CENTERS AND SHAPE OF MOLECULES 3915 03:13:39,452 --> 03:13:45,992 AND OTHER 3-D RELEVANT 3916 03:13:45,992 --> 03:13:46,326 PROPERTIES. 3917 03:13:46,326 --> 03:13:50,463 IN ORDER TO CONSIDER TO BE ABLE 3918 03:13:50,463 --> 03:13:53,033 TO COMPUTE 3 DIMENSIONAL 3919 03:13:53,033 --> 03:13:55,635 CONFIRMATIONS WE USED A 3920 03:13:55,635 --> 03:13:59,906 GRAPH-BASED REPRESENTATION OF 3921 03:13:59,906 --> 03:14:08,381 MOLECULES WITH VERTIX AND TYPES 3922 03:14:08,381 --> 03:14:10,383 TO REPRESENT AND COLOR THIS 3 3923 03:14:10,383 --> 03:14:11,151 DIMENSIONAL GRAPH AND TO 3924 03:14:11,151 --> 03:14:14,487 CONSIDER LOCAL NEIGHBORHOODS, 3925 03:14:14,487 --> 03:14:16,256 MESSAGE PASSING TURNED OUT TO BE 3926 03:14:16,256 --> 03:14:19,793 ESSENTIAL WHEN WE WISH TO WORK 3927 03:14:19,793 --> 03:14:22,796 WITH 3-D GRAPHS OF MOLECULAR 3928 03:14:22,796 --> 03:14:24,964 STRUCTURES AND CAN USE DEEP 3929 03:14:24,964 --> 03:14:28,468 NEURAL NETWORKS AND CASCADED 3930 03:14:28,468 --> 03:14:29,602 STACK NETWORKS TO COMPUTE ALL 3931 03:14:29,602 --> 03:14:32,806 KINDS OF PROPERTIES BASED ON 3932 03:14:32,806 --> 03:14:36,810 THESE UPDATED MOLECULAR FEATURES 3933 03:14:36,810 --> 03:14:37,477 OF THE MOLECULAR GRAPH. 3934 03:14:37,477 --> 03:14:38,878 THE FIRST EXAMPLE TO SHOWCASE 3935 03:14:38,878 --> 03:14:43,216 WHAT'S POSSIBLE WITH THE TOOLS 3936 03:14:43,216 --> 03:14:44,951 IS LEAD OPTIMIZATION. 3937 03:14:44,951 --> 03:14:48,455 WE GO FROM THE HIT TO LEAD STAGE 3938 03:14:48,455 --> 03:14:48,955 AND FUNCTIONALIZATION. 3939 03:14:48,955 --> 03:14:51,691 THE IDEA IS TO TAKE A MOLECULE 3940 03:14:51,691 --> 03:14:54,394 AND EXTEND TO GROW THIS MOLECULE 3941 03:14:54,394 --> 03:14:57,263 IN SOME CHEMICALLY MEANINGFUL 3942 03:14:57,263 --> 03:15:00,800 WAY SO IT CAN BE FUNCTIONALIZED 3943 03:15:00,800 --> 03:15:03,870 WITHOUT HAVING TO SYNTHESIZE THE 3944 03:15:03,870 --> 03:15:05,438 COMPLETE MOLECULE FROM SCRATCH. 3945 03:15:05,438 --> 03:15:08,808 IN ORDER TO DO SO WE STARTED 3946 03:15:08,808 --> 03:15:12,746 WITH THE VERSATILE CHEMICAL 3947 03:15:12,746 --> 03:15:14,948 REACTION WHICH IS FASHIONABLE 3948 03:15:14,948 --> 03:15:18,318 AND USEFUL IN CHEMISTRY WITH OUR 3949 03:15:18,318 --> 03:15:21,955 PROJECT PARTNERSES WE GENERATED 3950 03:15:21,955 --> 03:15:24,591 TRAINING DATA FOR THIS GRAPH 3951 03:15:24,591 --> 03:15:28,094 NEURAL NETWORK AND APPROXIMATELY 3952 03:15:28,094 --> 03:15:29,929 2,200 REACTIONS WERE PERFORMED 3953 03:15:29,929 --> 03:15:32,599 AND THIS IS A SMALL DATA SET BUT 3954 03:15:32,599 --> 03:15:34,200 WAS SUFFICIENT AND THE REACTION 3955 03:15:34,200 --> 03:15:35,902 EXAMPLES WERE USED IT TRAIN AN 3956 03:15:35,902 --> 03:15:41,808 EFFORT I WILL SHOW ON THE NEXT 3957 03:15:41,808 --> 03:15:49,082 SLIDE TO PREDICT POTENTIAL 3958 03:15:49,082 --> 03:15:51,384 BORLYATION SITES AND IF IT TOOK 3959 03:15:51,384 --> 03:15:52,819 PLACE, WHERE IN THE MOLECULE AND 3960 03:15:52,819 --> 03:16:03,129 WHAT'S THE YIELD. 3961 03:16:05,665 --> 03:16:12,705 AND WE HAVE THE 3-D OR 2-D GRAPH 3962 03:16:12,705 --> 03:16:16,309 USED AS INPUT AND FROM THE TWO 3963 03:16:16,309 --> 03:16:20,847 IMPORTANT STEPS ONE CAN COMPUTE 3964 03:16:20,847 --> 03:16:22,415 THESE DESIRED OUTPUTS. 3965 03:16:22,415 --> 03:16:28,221 UNUP PARTICULAR -- AND WE ADDED 3966 03:16:28,221 --> 03:16:30,323 A INPUT AND CHEMISTS CAN START 3967 03:16:30,323 --> 03:16:34,894 TO OWN THE DEEP LEARNING TOOLS 3968 03:16:34,894 --> 03:16:36,229 BY MODULATING THEIR 3969 03:16:36,229 --> 03:16:38,364 FUNCTIONALITY USING WELL 3970 03:16:38,364 --> 03:16:39,732 ESTABLISHED CHEMICAL KNOWLEDGE, 3971 03:16:39,732 --> 03:16:40,800 FOR EXAMPLE, REACTION 3972 03:16:40,800 --> 03:16:41,734 CONDITIONS. 3973 03:16:41,734 --> 03:16:44,070 THE OUTCOME WAS VERY SURPRISING. 3974 03:16:44,070 --> 03:16:46,906 WE HAD EXTREMELY GOOD SUCCESS 3975 03:16:46,906 --> 03:16:51,411 RATES, HIGH SUCCESS RATES FOR 3976 03:16:51,411 --> 03:16:52,812 THE BINARY PREDICTION WILL THE 3977 03:16:52,812 --> 03:16:57,150 REACTION TAKE PLACE, YES OR NO? 3978 03:16:57,150 --> 03:16:59,552 IF SO AND THE PREDICTIVE 3979 03:16:59,552 --> 03:17:01,287 REACTION YIELD. 3980 03:17:01,287 --> 03:17:04,257 THE ARROW PREDICTED IN THE ORDER 3981 03:17:04,257 --> 03:17:07,160 OF THE EXPERIMENTAL ARROW. 3982 03:17:07,160 --> 03:17:14,267 HERE ARE A FEW EXAMPLES TAKEN 3983 03:17:14,267 --> 03:17:16,703 FROM RAUSCH ONGOING PROJECTS IN 3984 03:17:16,703 --> 03:17:18,471 THEIR PHARMA DISCOVERY PROJECTS. 3985 03:17:18,471 --> 03:17:22,909 ON THE LEFT SIDE THE INPUT 3986 03:17:22,909 --> 03:17:26,579 MOLECULE AND ON THE RIGHT THE 3987 03:17:26,579 --> 03:17:27,914 RELATION PRODUCT AND YOU'LL SEE 3988 03:17:27,914 --> 03:17:31,217 IN THE MAJORITY OF THE CASES THE 3989 03:17:31,217 --> 03:17:32,819 PREDICTIONS OF SUCH A GRAPH 3990 03:17:32,819 --> 03:17:34,287 NEURAL NETWORK WERE CORRECT. 3991 03:17:34,287 --> 03:17:34,988 I'D LIKE TO HIGHLIGHT ONE 3992 03:17:34,988 --> 03:17:40,527 EXAMPLE THAT IS A FAILURE ON THE 3993 03:17:40,527 --> 03:17:41,828 LOWER RIGHT. 3994 03:17:41,828 --> 03:17:46,933 HERE ARE THE ALGORITHM PREDICTED 3995 03:17:46,933 --> 03:17:50,470 NO ATOM WOULD BE A SITE OF 3996 03:17:50,470 --> 03:17:51,437 CORRELATION BUT WE OBSERVED THE 3997 03:17:51,437 --> 03:17:55,842 CORRELATION AND WHY IS THAT? 3998 03:17:55,842 --> 03:17:59,679 THERE WAS NOT ONE EXAMPLE OF AN 3999 03:17:59,679 --> 03:18:02,482 CARBON ATOM AS A SUBSTRATE FOR 4000 03:18:02,482 --> 03:18:04,350 THE BORYLATION IN THE TRAINING 4001 03:18:04,350 --> 03:18:04,584 DATA. 4002 03:18:04,584 --> 03:18:07,387 SO THE SYSTEM CANNOT KNOW WHAT 4003 03:18:07,387 --> 03:18:08,488 IT HASN'T SEEN. 4004 03:18:08,488 --> 03:18:10,823 THIS IS SOMETHING ALSO TO KEEP 4005 03:18:10,823 --> 03:18:16,329 IN MIND WHEN IT COMES TO 4006 03:18:16,329 --> 03:18:19,399 TRAINING DATA SELECTION THAT WE 4007 03:18:19,399 --> 03:18:20,900 HAVE A RATHER DIVERSE AS 4008 03:18:20,900 --> 03:18:23,670 POSSIBLE COLLECTION OF INPUT 4009 03:18:23,670 --> 03:18:23,870 DATA. 4010 03:18:23,870 --> 03:18:25,972 IT'S NOT SO MUCH THE AMOUNT THE 4011 03:18:25,972 --> 03:18:28,508 NUMBER OF DATA POINTS WE USE FOR 4012 03:18:28,508 --> 03:18:32,812 MODEL TRAINING BUT THE DIVERSITY 4013 03:18:32,812 --> 03:18:41,321 OF VARIOUS INSTANCES. 4014 03:18:41,321 --> 03:18:43,289 LET'S COMBINE SEQUENCE-BASED 4015 03:18:43,289 --> 03:18:46,926 LANGUAGE MODEL TO DEVELOP THE 4016 03:18:46,926 --> 03:18:48,795 HOPEFULLY UNIVERSAL TOOL FOR 4017 03:18:48,795 --> 03:18:51,397 STRUCTURE BASED DESIGN SEQUENCE 4018 03:18:51,397 --> 03:18:54,400 RELATIONSHIP LEARNING. 4019 03:18:54,400 --> 03:18:56,502 AND IN ORDER TO DO SO WE NEED TO 4020 03:18:56,502 --> 03:18:58,705 CHANGE PERSPECTIVE WHEN IT COMES 4021 03:18:58,705 --> 03:18:59,339 TO HOW WE TRAIN THESE DEEP 4022 03:18:59,339 --> 03:19:02,742 ARCHITECTURES. 4023 03:19:02,742 --> 03:19:04,844 TRADITIONALLY, WE HAVE USED A 4024 03:19:04,844 --> 03:19:07,347 LIGAND CENTRIC VIEW. 4025 03:19:07,347 --> 03:19:10,583 IT CAN MAKE THE SPACE AND EACH 4026 03:19:10,583 --> 03:19:14,320 SMALL DOT MOLECULE A LIGAND AND 4027 03:19:14,320 --> 03:19:22,929 WANTED TO PREDICT A LIGAND BASED 4028 03:19:22,929 --> 03:19:24,530 DE NOVO DESIGN AND WE TRAINED A 4029 03:19:24,530 --> 03:19:28,134 NEURAL NETWORK ON THE 4030 03:19:28,134 --> 03:19:30,870 INTERACTION NETWORK OF PROTEINS 4031 03:19:30,870 --> 03:19:32,972 AND SMALL MOLECULAR LIGANDS AND 4032 03:19:32,972 --> 03:19:34,941 WE CAN NOW USE A GRAPH 4033 03:19:34,941 --> 03:19:40,146 REPRESENTATION OF THE BINDING 4034 03:19:40,146 --> 03:19:43,016 SITE AS INPUT AND LANGUAGE MODEL 4035 03:19:43,016 --> 03:19:46,619 AS DECODER SO THE CHEMICAL 4036 03:19:46,619 --> 03:19:47,153 LANGUAGE MODEL SUGGESTS 4037 03:19:47,153 --> 03:19:50,990 POTENTIAL LIGANDS FOR THE INPUT 4038 03:19:50,990 --> 03:19:55,728 LIGAND BINDING SITE AND POCKET. 4039 03:19:55,728 --> 03:19:57,096 WITH THAT APPROACH WE WENT INTO 4040 03:19:57,096 --> 03:20:02,502 THE LAB AND THIS IS THE VERY 4041 03:20:02,502 --> 03:20:04,804 FIRST EXAMPLE WE PUBLISHED ON 4042 03:20:04,804 --> 03:20:05,838 THE TOPIC. 4043 03:20:05,838 --> 03:20:09,175 STARTING FROM THE LIGAND BINDING 4044 03:20:09,175 --> 03:20:11,911 POCKET THAT WAS NOT IN THE 4045 03:20:11,911 --> 03:20:13,446 TRAINING DATA THE DE NOVO DESIGN 4046 03:20:13,446 --> 03:20:16,215 STRUCTURE GENERATOR SUGGESTS 4047 03:20:16,215 --> 03:20:18,117 MOLECULES 1 AND 2 HAVE 4048 03:20:18,117 --> 03:20:20,787 SYNTHESIZABLE IN THE LAB AND LAD 4049 03:20:20,787 --> 03:20:28,795 THE DESIRED ACTIVITY AND 4050 03:20:28,795 --> 03:20:34,167 SELECTIVITY. 4051 03:20:34,167 --> 03:20:36,602 AND THEY'RE NOT SUBSTRATES AND 4052 03:20:36,602 --> 03:20:43,910 THEY'RE SELECTIVE TO PFAR GAMMA. 4053 03:20:43,910 --> 03:20:48,648 THE LEARNING APPROACH IMPLICITLY 4054 03:20:48,648 --> 03:20:50,316 CONSIDERS LIGAND SELECTIVITY 4055 03:20:50,316 --> 03:20:51,951 WITHOUT HAVING TO EXPLICITLY 4056 03:20:51,951 --> 03:20:55,421 PREDICT ACTIVITY AND WE HAVE TO 4057 03:20:55,421 --> 03:20:58,758 SEE AS A COMMUNITY WHETHER THIS 4058 03:20:58,758 --> 03:21:02,895 APPROACH ENABLES LIGAND DESIGN 4059 03:21:02,895 --> 03:21:04,797 FOR ALL STRUCTURES AND ALPHA 4060 03:21:04,797 --> 03:21:06,599 FOLD PREDICTED STRUCTURES AND 4061 03:21:06,599 --> 03:21:10,403 THAT'S AN ONGOING ACTIVITY. 4062 03:21:10,403 --> 03:21:12,605 LET ME SUMMARIZE, RULE BASED DE 4063 03:21:12,605 --> 03:21:14,440 NOVO DESIGN PERFORMS WELL AND 4064 03:21:14,440 --> 03:21:20,813 RELIABLE WHEN YOU HAVE LITTLE 4065 03:21:20,813 --> 03:21:23,282 LIMITED, SCARCE DATA AVAILABLE 4066 03:21:23,282 --> 03:21:27,887 AND DEEP A.I. MODELS HAVE PROVEN 4067 03:21:27,887 --> 03:21:30,723 TO DELIVER BIO SYNTHESIZABLE 4068 03:21:30,723 --> 03:21:32,058 MOLECULES AND THE MOST CRITICAL 4069 03:21:32,058 --> 03:21:35,328 ASPECT OF DEEP LEARNING IN MY 4070 03:21:35,328 --> 03:21:36,395 VIEW IS PROBLEM REPRESENTATION 4071 03:21:36,395 --> 03:21:38,431 NOT SO MUCH THE TOTAL AMOUNT OF 4072 03:21:38,431 --> 03:21:39,832 DATA AND NOT SO MUCH THE CHOICE 4073 03:21:39,832 --> 03:21:44,837 OF A GENERATIVE A.I. BUT 4074 03:21:44,837 --> 03:21:48,407 TRAINING DATA SHAPING. 4075 03:21:48,407 --> 03:21:49,008 MOST IMPORTANTLY THERE'S NO 4076 03:21:49,008 --> 03:21:51,477 LEARNING WITHOUT FEEDBACK AND 4077 03:21:51,477 --> 03:21:52,845 FOR FEEDBACK WE NEED CHEMISTRY 4078 03:21:52,845 --> 03:21:55,081 BECAUSE WE NEED TO MAKE 4079 03:21:55,081 --> 03:21:56,816 MOLECULES AS A COMMUNITY, 4080 03:21:56,816 --> 03:21:57,817 SYNTHESIZE, TEST THEM. 4081 03:21:57,817 --> 03:22:02,121 AS A COMMUNITY I WOULD SAY TEAM 4082 03:22:02,121 --> 03:22:04,624 UP CHEMISTS WITH COMPUTER 4083 03:22:04,624 --> 03:22:07,126 SCIENTISTS, WITH BIOLOGISTS AND 4084 03:22:07,126 --> 03:22:10,029 VICE VERSA BECAUSE WE NEED MUCH 4085 03:22:10,029 --> 03:22:11,531 MORE CHEMISTRY. 4086 03:22:11,531 --> 03:22:12,231 WITH EVERYTHING WE'RE DOING IN 4087 03:22:12,231 --> 03:22:13,966 THE FASCINATING FIELD AREA YOU 4088 03:22:13,966 --> 03:22:16,636 SHOULD NOT FORGET WE CANNOT 4089 03:22:16,636 --> 03:22:20,573 ESCAPE THE EDGE OF CHAOS AND THE 4090 03:22:20,573 --> 03:22:21,374 PARTIAL PREDICTABILITY AND CAN 4091 03:22:21,374 --> 03:22:24,076 PUSH THE ENVELOPE TOGETHER. 4092 03:22:24,076 --> 03:22:26,612 I THANK TONS OF MY CO-WORKERS 4093 03:22:26,612 --> 03:22:27,814 THAT CONTRIBUTED TO THE 4094 03:22:27,814 --> 03:22:29,782 DEVELOPMENTS OVER THE LAST 30 4095 03:22:29,782 --> 03:22:30,983 YEARS. 4096 03:22:30,983 --> 03:22:31,918 MY COLLABORATORS AND SPONSORS 4097 03:22:31,918 --> 03:22:34,887 AND YOU FOR LISTENING AND I'LL 4098 03:22:34,887 --> 03:22:36,122 BE HAPPY TO ANSWER ANY QUESTION 4099 03:22:36,122 --> 03:22:37,023 YOU MAY HAVE. 4100 03:22:37,023 --> 03:22:40,059 >> THANK YOU FOR A VERY 4101 03:22:40,059 --> 03:22:40,927 STRUCTURED TALK. 4102 03:22:40,927 --> 03:22:43,896 WE ARE NOT GOING TO HAVE TOO 4103 03:22:43,896 --> 03:22:45,398 MUCH TIME FOR QUESTIONS. 4104 03:22:45,398 --> 03:22:48,134 KINDLY ANSWER WHAT YOU SEE ON 4105 03:22:48,134 --> 03:22:48,835 THE SCREEN AFTER. 4106 03:22:48,835 --> 03:22:51,771 BUT A COUPLE QUESTIONS I'D LIKE 4107 03:22:51,771 --> 03:22:56,042 TO ANSWER QUICKLY, GIVEN THE 4108 03:22:56,042 --> 03:22:58,110 SCARCE DATA PROBLEM AND WITH 4109 03:22:58,110 --> 03:22:59,979 ALPHA FOLD FILLING IN THE GAP 4110 03:22:59,979 --> 03:23:01,647 WHAT KIND OF EXPERIMENTAL DATA 4111 03:23:01,647 --> 03:23:08,521 ARE NOW MOST CRITICALLY IN NEED? 4112 03:23:08,521 --> 03:23:10,923 >> IN MY VIEW IT'S STILL 4113 03:23:10,923 --> 03:23:16,796 ACTIVITY DATA FOR THE BROADER 4114 03:23:16,796 --> 03:23:19,165 DIVERSITY OF TARGETS. 4115 03:23:19,165 --> 03:23:21,434 >> OKAY. 4116 03:23:21,434 --> 03:23:25,371 OFF-TARGET ATTEMPT, HOW DO YOU 4117 03:23:25,371 --> 03:23:25,605 DESIGN? 4118 03:23:25,605 --> 03:23:27,073 >> OFF TARGETS? 4119 03:23:27,073 --> 03:23:32,078 WELL, FOR A LONG TIME WE'VE 4120 03:23:32,078 --> 03:23:34,714 TRIED TO AND HAVE DEVELOPED 4121 03:23:34,714 --> 03:23:39,318 INDIVIDUAL PREDICTION TOOLS FOR 4122 03:23:39,318 --> 03:23:41,187 EACH TARGET ONE ACTIVITY AFTER 4123 03:23:41,187 --> 03:23:48,794 THE OTHER AND MULTITASKING AND 4124 03:23:48,794 --> 03:23:53,532 TO PREDICT AS YES/NO DESIGN 4125 03:23:53,532 --> 03:23:55,668 DECISIONS OR BINDS OR DOESN'T 4126 03:23:55,668 --> 03:23:56,969 BINE AND WITH THE INTERACTON 4127 03:23:56,969 --> 03:23:59,906 LEARNING APPROACH WE TRIED TO 4128 03:23:59,906 --> 03:24:03,276 CAPTURE ACTIVITY AND SELECTIVELY 4129 03:24:03,276 --> 03:24:04,310 IMPLICITLY RATHER THAN 4130 03:24:04,310 --> 03:24:05,044 EXPLICITLY BY TRAINING THE 4131 03:24:05,044 --> 03:24:06,178 NEURAL NETWORK OR DEEP LEARNING 4132 03:24:06,178 --> 03:24:07,313 MODEL ON THE NETWORK OF 4133 03:24:07,313 --> 03:24:12,051 INTERACTIONS. 4134 03:24:12,051 --> 03:24:13,286 SO, ONLY RELEVANT INTERACTIONS 4135 03:24:13,286 --> 03:24:23,829 LIGANDED -- LIGANDS AND TARGETS 4136 03:24:24,697 --> 03:24:25,898 ARE SELECTED. 4137 03:24:25,898 --> 03:24:26,632 >> THANK YOU. 4138 03:24:26,632 --> 03:24:28,634 WE'LL HAVE TO CLOSE THE SESSION 4139 03:24:28,634 --> 03:24:30,236 AND KINDLY ANSWER QUESTIONS WE 4140 03:24:30,236 --> 03:24:31,304 DIDN'T HAVE TIME FOR. 4141 03:24:31,304 --> 03:24:35,308 LET'S MOVE ON WITH THE SESSION. 4142 03:24:35,308 --> 03:24:38,978 THE NEXT SPEAKER IS PROFESSOR 4143 03:24:38,978 --> 03:24:43,382 CON 4144 03:24:43,382 --> 03:24:44,817 CONNOR COLEY AT MIT IN THE 4145 03:24:44,817 --> 03:24:47,086 DEPARTMENT OF CHEMICAL 4146 03:24:47,086 --> 03:24:50,189 ENGINEERING AND DEPARTMENT OF 4147 03:24:50,189 --> 03:24:51,490 ELECTRICAL ENGINEERING AND 4148 03:24:51,490 --> 03:24:54,093 COMPUTER SCIENCE AND HIS GROUP 4149 03:24:54,093 --> 03:24:56,329 WORKS WITH MODELS TO UNDERSTAND 4150 03:24:56,329 --> 03:25:01,267 HOW MOLECULES BEHAVE, INTERACT 4151 03:25:01,267 --> 03:25:07,540 AND REACT AND USES THE 4152 03:25:07,540 --> 03:25:09,842 INFORMATION FOR DISCOVERY. 4153 03:25:09,842 --> 03:25:20,386 CONNOR HAS RECEIVED REMARKABLE 4154 03:25:20,853 --> 03:25:26,492 AWARDS. 4155 03:25:26,492 --> 03:25:27,827 WE LOOK FORWARD TO HEARING FROM 4156 03:25:27,827 --> 03:25:28,160 YOU. 4157 03:25:28,160 --> 03:25:28,928 >> THANK YOU FOR ORGANIZING THE 4158 03:25:28,928 --> 03:25:31,063 SESSION AND FOR THE INVITATION 4159 03:25:31,063 --> 03:25:34,700 TO SHARE SOME OF OUR WORK WE'RE 4160 03:25:34,700 --> 03:25:38,337 DOING HERE AT MIT. 4161 03:25:38,337 --> 03:25:41,007 I WANT TO START THE TALK TO 4162 03:25:41,007 --> 03:25:43,676 AGGRESSIVELY AGREE WITH MANY 4163 03:25:43,676 --> 03:25:46,912 THINGS WE JUST HEARD FROM 4164 03:25:46,912 --> 03:25:47,980 GISBERT SCHNEIDER ABOUT 4165 03:25:47,980 --> 03:25:49,749 MOLECULAR DESIGN AND THE 4166 03:25:49,749 --> 03:25:52,518 CONSIDERATION OF SYNTHESIS. 4167 03:25:52,518 --> 03:25:53,753 THAT WILL BE A THEME OF THIS 4168 03:25:53,753 --> 03:25:54,387 TALK FOR THE NEXT 25 MINUTES OR 4169 03:25:54,387 --> 03:25:57,289 SO. 4170 03:25:57,289 --> 03:26:01,093 IF WE WANT TO THINK ABOUT THE 4171 03:26:01,093 --> 03:26:04,397 GOAL OF MOLECULAR DESIGN 4172 03:26:04,397 --> 03:26:07,867 OPTIMIZATION WE CAN TRY TO MAP 4173 03:26:07,867 --> 03:26:09,835 IT INTO A CLASSICAL FORMULATION 4174 03:26:09,835 --> 03:26:11,704 TRYING TO FIND OR GIVEN FUNCTION 4175 03:26:11,704 --> 03:26:13,172 THAT RETURNS PERHAPS AN ACTIVITY 4176 03:26:13,172 --> 03:26:16,042 OR SOME MPO SCORE. 4177 03:26:16,042 --> 03:26:18,411 THE BEST DESIGN FROM THIS LARGE 4178 03:26:18,411 --> 03:26:20,813 SPACE OF ALL POSSIBLE MOLECULES 4179 03:26:20,813 --> 03:26:24,817 AND WE MIGHT THEN CONSIDER THE 4180 03:26:24,817 --> 03:26:26,485 OPTIMAL MOLECULE WHAT WE WOULD 4181 03:26:26,485 --> 03:26:28,320 ADVANCE FURTHER INTO TESTING. 4182 03:26:28,320 --> 03:26:33,292 IN REALITY OFTEN WE WANT TO 4183 03:26:33,292 --> 03:26:36,395 OPTIM 4184 03:26:36,395 --> 03:26:36,929 OPTIM 4185 03:26:36,929 --> 03:26:38,798 OPTIMIZE PROPERTIES DIFFICULT 4186 03:26:38,798 --> 03:26:40,800 TO MEASURE WITH A MEASUREMENT 4187 03:26:40,800 --> 03:26:42,535 THAT APPROXIMATES SOMETHING WE'D 4188 03:26:42,535 --> 03:26:44,303 LIKE TO OPTIMIZE FOR. 4189 03:26:44,303 --> 03:26:46,906 IT MATE LOOK LIKE BINDING VERSUS 4190 03:26:46,906 --> 03:26:48,140 IN IN VIVO EFFICACY. 4191 03:26:48,140 --> 03:26:50,076 BINDING BEING THE PROXY FOR THE 4192 03:26:50,076 --> 03:26:50,376 OTHER. 4193 03:26:50,376 --> 03:26:52,812 IN REALITY, PERHAPS IT'S NOT 4194 03:26:52,812 --> 03:26:55,081 TRULY THE SPACE OF ALL POSSIBLE 4195 03:26:55,081 --> 03:26:56,549 MOLECULES WE'RE TRYING TO 4196 03:26:56,549 --> 03:26:58,918 NAVIGATE BUT THE SPACE OF 4197 03:26:58,918 --> 03:26:59,919 SYNTHESISABLE MOLECULES BECAUSE 4198 03:26:59,919 --> 03:27:01,787 THESE ARE THE ONES WE CAN'T TAKE 4199 03:27:01,787 --> 03:27:03,255 IN THE LABORATORY TO PHYSICALLY 4200 03:27:03,255 --> 03:27:08,794 TEST AND USE FOR FEEDBACK. 4201 03:27:08,794 --> 03:27:10,763 SO I VISUALIZE THE LANDSCAPE 4202 03:27:10,763 --> 03:27:13,933 THAT CONNECTS STRUCTURE ON THE 4203 03:27:13,933 --> 03:27:16,702 HORIZONTAL PLANE HERE TO THE 4204 03:27:16,702 --> 03:27:20,806 PROPERTIES THAT EXHIBIT SHOWN AS 4205 03:27:20,806 --> 03:27:22,908 PEAKS AND TROUGHS. 4206 03:27:22,908 --> 03:27:27,513 THE GOAL OF ELECTED DISCOVERY IS 4207 03:27:27,513 --> 03:27:29,181 THE MOLECULES THAT EXIST AT 4208 03:27:29,181 --> 03:27:39,625 THESE LOCAL MAXIMA PEAKS. 4209 03:27:47,500 --> 03:27:51,770 WHILE WE'VE HAD THIS TESTING OF 4210 03:27:51,770 --> 03:27:53,372 MOLECULE FROM A VAST DESIGN 4211 03:27:53,372 --> 03:27:55,441 SPACE WE CAN USE TOPOGRAPH CAL 4212 03:27:55,441 --> 03:27:56,809 MODELS TO STEER US IN THE RIGHT 4213 03:27:56,809 --> 03:28:00,813 DIRECTION TO UNDERSTAND WHERE WE 4214 03:28:00,813 --> 03:28:05,818 SHOULD BE LOOKING FOR NEW 4215 03:28:05,818 --> 03:28:08,287 MOLECULE AND WAYS TO LOOK RAT 4216 03:28:08,287 --> 03:28:16,362 AGAIN -- GENERATIVE DESIGN OR 4217 03:28:16,362 --> 03:28:17,530 DE NOVO DESIGN AND USE 4218 03:28:17,530 --> 03:28:18,964 GENERATIVE MODELS TO HELP CLIMB 4219 03:28:18,964 --> 03:28:22,968 THE PEAKS AND GET TO THESE 4220 03:28:22,968 --> 03:28:27,973 MAXIMA MER QUICKLY. 4221 03:28:27,973 --> 03:28:28,641 WE'LL BE THINKING ABOUT THE 4222 03:28:28,641 --> 03:28:31,477 DESIGN ASPECT IN THE REST OF THE 4223 03:28:31,477 --> 03:28:36,182 TALK AND MEANT TO FEED INTO 4224 03:28:36,182 --> 03:28:37,683 SYNTHESIS AND TESTING. 4225 03:28:37,683 --> 03:28:43,355 WE WON'T EVER HAVE THE FIDELITY 4226 03:28:43,355 --> 03:28:46,625 TO PREDICT IN SILICO FOR SOME 4227 03:28:46,625 --> 03:28:49,295 TIME SO WE'RE RELIANT ON THE 4228 03:28:49,295 --> 03:28:52,798 EXPERIMENTAL SPEED BACKTO 4229 03:28:52,798 --> 03:28:54,366 RESTRREVISE 4230 03:28:54,366 --> 03:28:56,802 OR MODELS AND DESIGNS OF FOR THE 4231 03:28:56,802 --> 03:28:57,703 NEXT ITERATION. 4232 03:28:57,703 --> 03:29:02,208 ALL THE MODELS ARE FRAMED IN 4233 03:29:02,208 --> 03:29:05,611 THIS ITERATIVE DRUG DISCOVERY 4234 03:29:05,611 --> 03:29:05,878 VIEW. 4235 03:29:05,878 --> 03:29:10,950 SO WHAT ARE THE CORNERSTONES OF 4236 03:29:10,950 --> 03:29:14,653 DISCOVERY CHEMISTRY SETTINGS HAS 4237 03:29:14,653 --> 03:29:16,922 TO BE PARALLEL PLATE-BASED 4238 03:29:16,922 --> 03:29:17,523 CHEMISTRY. 4239 03:29:17,523 --> 03:29:19,558 GIVEN SOME COMMON BUILDING 4240 03:29:19,558 --> 03:29:22,928 BLOCKS OR MOTIFS ONE CAN 4241 03:29:22,928 --> 03:29:26,398 CONSTRUCT RAPIDLY 96 OR 384 OR 4242 03:29:26,398 --> 03:29:28,167 SOME NUMBER OF DERIVATIVES. 4243 03:29:28,167 --> 03:29:30,135 IN THE CASE WE SAW FROM 4244 03:29:30,135 --> 03:29:34,473 PROFESSOR SCHNEIDER WAS IT WAS 4245 03:29:34,473 --> 03:29:35,908 CONCEPTUAL FOR LATE STAGE AND 4246 03:29:35,908 --> 03:29:38,210 OTHERS FOR COUPLING OF DIFFERENT 4247 03:29:38,210 --> 03:29:40,813 BUILDING BLOCKS IN THIS CASE 4248 03:29:40,813 --> 03:29:41,680 THROUGH A COUPLING. 4249 03:29:41,680 --> 03:29:43,849 AND SO THIS IS SUMMARIZING 4250 03:29:43,849 --> 03:29:45,217 COLLABORATIVE WORK DONE WITH 4251 03:29:45,217 --> 03:29:47,519 ABBVIE WHERE WE LOOKED AT THE 4252 03:29:47,519 --> 03:29:49,755 HISTORICAL DATA ON COUPLINGS 4253 03:29:49,755 --> 03:29:52,558 FROM THE MEDICINAL CHEMISTRY 4254 03:29:52,558 --> 03:29:53,692 SETTINGS TO TRAIN REACTION 4255 03:29:53,692 --> 03:29:55,327 YIELDS AND UNDERSTAND THE 4256 03:29:55,327 --> 03:30:00,766 LIKELIHOODS OF SUCCESSI AND TAK 4257 03:30:00,766 --> 03:30:04,803 THE DESIGNING LIBRARIES COUPLING 4258 03:30:04,803 --> 03:30:09,375 ONE WITH MANY ACIDS AND DECIDE 4259 03:30:09,375 --> 03:30:11,543 WELL, A CHEMIST DESIGN THIS 4260 03:30:11,543 --> 03:30:15,481 LIBRARY WITH A CERTAIN SELECTION 4261 03:30:15,481 --> 03:30:19,051 OF 46 BRONNIC ACIDS AND IF WE 4262 03:30:19,051 --> 03:30:21,086 FOLLOWED THOSE SELECTIONS 4263 03:30:21,086 --> 03:30:24,423 EXACTLY, OF THE 46 DESIGNED BY 4264 03:30:24,423 --> 03:30:29,728 HUMAN CHEMISTS, 9 6 SUCCEEDED TO 4265 03:30:29,728 --> 03:30:31,497 BE TESTED FOR BIOLOGICAL 4266 03:30:31,497 --> 03:30:32,298 ACTIVITY. 4267 03:30:32,298 --> 03:30:34,400 BY INSERTING THE MODEL THAT 4268 03:30:34,400 --> 03:30:39,004 THINK OF CHEMICAL YIELDS WE WERE 4269 03:30:39,004 --> 03:30:40,005 ABLE TO DETECT SOME OF THOSE 4270 03:30:40,005 --> 03:30:43,275 WOULD BE UNSUITABLE AS A 4271 03:30:43,275 --> 03:30:44,476 COUPLING MODEL FOR THIS MOTIF 4272 03:30:44,476 --> 03:30:46,211 AND COULD REPLACE SOME OF THOSE 4273 03:30:46,211 --> 03:30:51,183 DECISIONS AND END UP INCREASING 4274 03:30:51,183 --> 03:30:52,785 MODESTLY THE SUCCESS RATE. 4275 03:30:52,785 --> 03:30:55,954 AND FIVE ADDITIONAL COMPOUNDS 4276 03:30:55,954 --> 03:30:59,224 WERE ABLE TO BE MADE AND CARRIED 4277 03:30:59,224 --> 03:31:00,793 FORWARD TO PURIFICATION. 4278 03:31:00,793 --> 03:31:02,795 THIS IS A SINGLE STEP LIBRARY 4279 03:31:02,795 --> 03:31:04,363 AND WE'RE OFTEN INTERESTED IN 4280 03:31:04,363 --> 03:31:08,801 BUILDING UP LIBRARIES FROM 4281 03:31:08,801 --> 03:31:10,936 MULTIPLE SYNTHETIC STEPS AND 4282 03:31:10,936 --> 03:31:15,074 GOES TO REACTION TEMPLATE OR 4283 03:31:15,074 --> 03:31:16,775 STEP CONSTRAIN DESIGN OF NEW 4284 03:31:16,775 --> 03:31:17,042 MOLECULES. 4285 03:31:17,042 --> 03:31:19,378 ONE OF THE STRATEGIES WE'VE 4286 03:31:19,378 --> 03:31:22,948 TAKEN TO DEVELOPING LIBRARIES OF 4287 03:31:22,948 --> 03:31:24,283 MOLECULES CENTERED AROUND 4288 03:31:24,283 --> 03:31:28,620 PERHAPS KNOWN SEEDS OR HIP 4289 03:31:28,620 --> 03:31:32,791 COMPOUNDS WE'D LIKE TO EXPAND 4290 03:31:32,791 --> 03:31:34,927 AND TAKE A MOLECULE THROUGH 4291 03:31:34,927 --> 03:31:36,795 SYNTHESIS AND GIVEN THE PATHWAY 4292 03:31:36,795 --> 03:31:39,598 SPREAD A MAP OUT ALL THE ANALYZE 4293 03:31:39,598 --> 03:31:41,734 WE CAN BELIEVE WE CAN ACCESS BY 4294 03:31:41,734 --> 03:31:43,902 SWAPPING BUILDING BLOCKS FOR 4295 03:31:43,902 --> 03:31:44,670 EACH OTHER. 4296 03:31:44,670 --> 03:31:46,505 AND WE SAW WHAT KINDS OF 4297 03:31:46,505 --> 03:31:47,406 REACTIONS WE'LL USE TO BRING 4298 03:31:47,406 --> 03:31:49,641 TOGETHER THE COMPONENTS TO 4299 03:31:49,641 --> 03:31:51,777 PRODUCE OUR PRODUCTS AND TRYING 4300 03:31:51,777 --> 03:31:53,779 TO SIMPLY EXCHANGE ONE MOLECULE 4301 03:31:53,779 --> 03:31:57,983 FOR ANOTHER BUT STILL FOLLOW THE 4302 03:31:57,983 --> 03:31:59,251 SAME SYNTHETIC STRATEGY. 4303 03:31:59,251 --> 03:32:07,926 GIVEN THE PATHWAY WE RELY ON THE 4304 03:32:07,926 --> 03:32:09,495 TRANSFORMATION RULES AND WITH 4305 03:32:09,495 --> 03:32:11,330 SIMPLE INFORMATICS WE CAN 4306 03:32:11,330 --> 03:32:12,831 QUICKLY MAP OUT SOMETIMES 4307 03:32:12,831 --> 03:32:18,437 HUNDREDS OF ANALOGS, SOMETIMES 4308 03:32:18,437 --> 03:32:20,572 MILLIONS OF CERTAIN MOLECULAR 4309 03:32:20,572 --> 03:32:20,873 STRUCTURES. 4310 03:32:20,873 --> 03:32:22,941 WE'RE NOT TRYING TO BE CREATIVE 4311 03:32:22,941 --> 03:32:29,214 IN OUR APPROACH BUT TRYING TO 4312 03:32:29,214 --> 03:32:31,717 REPLICATE THE SYNTHETIC STRATEGY 4313 03:32:31,717 --> 03:32:33,385 TO CHANGE THE MOLECULES. 4314 03:32:33,385 --> 03:32:36,321 WHEN WE APPLIED THIS IN PRACTICE 4315 03:32:36,321 --> 03:32:38,357 FOCUSSING ON A COMPUTATIONAL 4316 03:32:38,357 --> 03:32:40,392 EVALUATION, WE'VE SHOWN YOU CAN 4317 03:32:40,392 --> 03:32:43,829 FIRST DO A VIRTUAL SCREEN FROM A 4318 03:32:43,829 --> 03:32:44,830 FIXED ENUMERATED LIBRARY LIKE 4319 03:32:44,830 --> 03:32:48,400 250,000 MOLECULES FROM ZINC HAS 4320 03:32:48,400 --> 03:32:51,937 BECOME A COMMON SUBSET TO USE 4321 03:32:51,937 --> 03:32:54,406 FOR APPROACHES AND WE FOUND ITS 4322 03:32:54,406 --> 03:32:56,542 PROPERTY SCORE WAS JUST UNDER 4323 03:32:56,542 --> 03:32:56,809 0.7. 4324 03:32:56,809 --> 03:32:58,744 THIS IS A NORMALIZED PROPERTY 4325 03:32:58,744 --> 03:33:00,512 SCORE WE'RE TRYING TO MAXIMIZE. 4326 03:33:00,512 --> 03:33:01,914 AND THROUGH THIS ENUMERATION 4327 03:33:01,914 --> 03:33:04,183 APPROACH WE CAN EXPAND OUT A 4328 03:33:04,183 --> 03:33:06,752 SPACE OF SYNTHESIZABLE ANALOGS 4329 03:33:06,752 --> 03:33:08,754 THAT ENDS UP INCLUDING THOUSANDS 4330 03:33:08,754 --> 03:33:10,489 OF NOVEL STRUCTURES WITH 4331 03:33:10,489 --> 03:33:11,423 PROPERTIES THAT EXCEED THE 4332 03:33:11,423 --> 03:33:12,291 ORIGINAL HIT. 4333 03:33:12,291 --> 03:33:14,326 WE CAN TAKE A STARTING POINT OR 4334 03:33:14,326 --> 03:33:16,395 IDEA FOR A MOLECULE AND RAPIDLY 4335 03:33:16,395 --> 03:33:20,232 EXPAND IT WITHOUT TRYING TO BE 4336 03:33:20,232 --> 03:33:20,833 CREATIVE ABOUT ITS CHEMICAL 4337 03:33:20,833 --> 03:33:22,968 SYNTHESIS AND STILL FIND NEW 4338 03:33:22,968 --> 03:33:23,602 STRUCTURES THAT MAY EXCEED ITS 4339 03:33:23,602 --> 03:33:27,806 PROPERTIES. 4340 03:33:27,806 --> 03:33:30,909 THIS STRATEGY IS FOLLOWING A 4341 03:33:30,909 --> 03:33:32,377 SINGLE PATHWAY AND IS RELYING 4342 03:33:32,377 --> 03:33:37,249 SOLELY ON THE IDEAS OF 4343 03:33:37,249 --> 03:33:39,117 ENUMERATION GROUNDED IN 4344 03:33:39,117 --> 03:33:39,418 INFORMATICS. 4345 03:33:39,418 --> 03:33:41,920 WE'RE ALSO INTERESTED IN THE USE 4346 03:33:41,920 --> 03:33:43,121 OF GENERATIVE OR DE NOVO DESIGN 4347 03:33:43,121 --> 03:33:48,093 MODELS TO INVENT NEW STRUCTURES. 4348 03:33:48,093 --> 03:33:50,262 THEY CAN BE VERY CREATIVE AND 4349 03:33:50,262 --> 03:33:51,363 COME UP WITH NEW DESIGNS WE 4350 03:33:51,363 --> 03:33:53,799 DIDN'T THINK OF AND PERHAPS 4351 03:33:53,799 --> 03:33:59,738 UNLOCK NEW CHEMICAL SPACE THAT 4352 03:33:59,738 --> 03:34:01,106 GO BEYOND 42 BILLIONS OF 4353 03:34:01,106 --> 03:34:04,009 STRUCTURES WE FIND IN LIBRARIES. 4354 03:34:04,009 --> 03:34:07,713 THE DOWN SIDE OF USING THESE 4355 03:34:07,713 --> 03:34:10,148 GENERATIVE TECHNIQUES AT TIMES 4356 03:34:10,148 --> 03:34:13,085 THEY CAN END UP BEING CREATIVE 4357 03:34:13,085 --> 03:34:13,852 TO A FAULT. 4358 03:34:13,852 --> 03:34:16,421 THE CREATIVITY REFLECTED ON THE 4359 03:34:16,421 --> 03:34:17,956 SLIDE WITH SOME STRUCTURES 4360 03:34:17,956 --> 03:34:22,494 MANIFESTS IN IDEAS THAT ARE VERY 4361 03:34:22,494 --> 03:34:26,732 NONSENSICAL TO PUT IT LIGHTLY, 4362 03:34:26,732 --> 03:34:28,400 BECAUSE THEY'RE EITHER NOT 4363 03:34:28,400 --> 03:34:31,036 STABLE OR NOT SYNTHESIZABLE AND 4364 03:34:31,036 --> 03:34:31,937 NOT REASONABLE STRUCTURES YOU'D 4365 03:34:31,937 --> 03:34:33,572 WANT TO TAKE IN THE LABORATORY 4366 03:34:33,572 --> 03:34:36,542 OR SUBMIT TO A CONTRACT 4367 03:34:36,542 --> 03:34:38,176 ORGANIZATION FOR SYNTHESIS OR 4368 03:34:38,176 --> 03:34:40,646 SUBMIT TO YOUR ROBOTIC PLATFORM 4369 03:34:40,646 --> 03:34:42,514 FOR SYNTHESIS IF THAT'S THE WAY 4370 03:34:42,514 --> 03:34:43,015 YOU OPERATE. 4371 03:34:43,015 --> 03:34:44,416 SO WE'RE INTERESTED IN 4372 03:34:44,416 --> 03:34:47,486 GENERATIVE DESIGN BUT WANT TO 4373 03:34:47,486 --> 03:34:50,923 AVOID THIS FAIRLY CATASTROPHIC 4374 03:34:50,923 --> 03:34:51,390 FAILURE MODE. 4375 03:34:51,390 --> 03:34:54,226 SO WHAT IF WE WANT TO USE 4376 03:34:54,226 --> 03:34:55,360 GENERATIVE DESIGN BUT STILL WANT 4377 03:34:55,360 --> 03:34:56,161 TO RESPECT SYNTHESIS? 4378 03:34:56,161 --> 03:35:00,799 THIS IS WHAT TAKES US BACK FROM 4379 03:35:00,799 --> 03:35:04,002 THE IDEATION THAT WE CAN FOLLOW 4380 03:35:04,002 --> 03:35:08,540 TO THIS STEP BY STEP REACTION 4381 03:35:08,540 --> 03:35:09,007 CONSTRAINED APPROACH. 4382 03:35:09,007 --> 03:35:12,811 SO IT'S MOST COMMON TO USE 4383 03:35:12,811 --> 03:35:14,479 GENERATIVE DESIGN FOR MOLECULAR 4384 03:35:14,479 --> 03:35:16,515 STRUCTURES IN WAYS THAT BUILD UP 4385 03:35:16,515 --> 03:35:19,585 NEW MOLECULES LETTER BY LETTER 4386 03:35:19,585 --> 03:35:22,387 AND WITH GRAPHS AND FRAGMENT OR 4387 03:35:22,387 --> 03:35:30,128 FRAGMENT OR ATOM BY ATOM WITH 3 4388 03:35:30,128 --> 03:35:31,563 DEM DIMENSIONAL MOLECULES BUT 4389 03:35:31,563 --> 03:35:33,198 YOU CAN END UP WITH DESIGNS THAT 4390 03:35:33,198 --> 03:35:40,806 ARE OKAY ACCORDING TO TOOLS BUT 4391 03:35:40,806 --> 03:35:51,283 NOT REALISTIC OR SYNTHETIC 4392 03:35:52,718 --> 03:35:56,788 TRACTABLE AND LOOKING THROUGH 4393 03:35:56,788 --> 03:36:01,593 THE LENS OF DEEP LEARNING AND 4394 03:36:01,593 --> 03:36:03,195 MODELS TO DESIGN STUCK TOURS NOT 4395 03:36:03,195 --> 03:36:07,699 AS LETTERS OR ATOMS BUT BUILDING 4396 03:36:07,699 --> 03:36:11,737 UP A REACTION PATHWAY. 4397 03:36:11,737 --> 03:36:14,806 AN EARLY CONTRIBUTION WAS A 4398 03:36:14,806 --> 03:36:18,377 MODEL TRAINING A NEURAL NETWORK 4399 03:36:18,377 --> 03:36:21,813 POLICY THAT COULD SELECT 4400 03:36:21,813 --> 03:36:22,948 COMMERCIALLY AVAILABLE BUILDING 4401 03:36:22,948 --> 03:36:26,685 BLOCKS AND TEMPLATES TO COMPOSE 4402 03:36:26,685 --> 03:36:28,253 A SYNTHETIC TREE FROM THE BOTTOM 4403 03:36:28,253 --> 03:36:28,420 UP. 4404 03:36:28,420 --> 03:36:32,891 WE CAN GIVE IT TO ACCESS OF 4405 03:36:32,891 --> 03:36:35,460 BUILDING BLOCKS AND 100 EXPERT 4406 03:36:35,460 --> 03:36:38,697 DEFINED REACTION TEMPLATES. 4407 03:36:38,697 --> 03:36:40,098 AND IT WILL COMPOSE WHATEVER 4408 03:36:40,098 --> 03:36:43,335 PATHWAYS IT WANTS WITH THE GOAL 4409 03:36:43,335 --> 03:36:45,303 OF REACHING A MOLECULE THAT 4410 03:36:45,303 --> 03:36:46,938 ACHIEVE A PROPERTY TARGET. 4411 03:36:46,938 --> 03:36:50,942 WE CAN TRAIN THE POLICY FOR 4412 03:36:50,942 --> 03:36:54,012 MODELS WE DO KNOW HOW TO MAKE 4413 03:36:54,012 --> 03:36:55,514 AND THEN OPTIMIZE WITH THOSE 4414 03:36:55,514 --> 03:36:57,783 WE'RE TRYING TO OPTIMIZE. 4415 03:36:57,783 --> 03:36:59,017 WE CAN CHANGE CONDITIONAL INPUT 4416 03:36:59,017 --> 03:37:04,790 TO THE MODEL THAT WILL GIVE US A 4417 03:37:04,790 --> 03:37:08,260 DIFFERENT SYNTHETIC PRIMARY 4418 03:37:08,260 --> 03:37:10,529 STRUCTURE AND THIS LOOP IS THEN 4419 03:37:10,529 --> 03:37:13,098 CONDUCIVE TO OUR OPTIMIZATION OR 4420 03:37:13,098 --> 03:37:15,367 ANY MODEL-FREE OPTIMIZATION 4421 03:37:15,367 --> 03:37:16,468 APPROACH WE WOULD LIKE. 4422 03:37:16,468 --> 03:37:18,503 AS A CONSEQUENCE OF TRAINING THE 4423 03:37:18,503 --> 03:37:21,540 GENERATIVE MODELS THAT THINK IN 4424 03:37:21,540 --> 03:37:23,341 TERMS OF SYNTHESIS RATHER THAN 4425 03:37:23,341 --> 03:37:26,445 ATOMS OR FRAGMENTS WE ENDS UP 4426 03:37:26,445 --> 03:37:29,147 FINDING SOLUTIONS WHERE THE 4427 03:37:29,147 --> 03:37:34,219 STRUCTURES ARE FAR SIM POLICE 4428 03:37:34,219 --> 03:37:41,059 OFFICER AND MORE SYNTHETICALLY 4429 03:37:41,059 --> 03:37:43,662 TRA 4430 03:37:43,662 --> 03:37:44,329 TRACTABLE AND DIFFICULT TO BRING 4431 03:37:44,329 --> 03:37:46,998 TO THE LABORATORY TO TEST. 4432 03:37:46,998 --> 03:37:48,467 USING THE COMMERCIAL BLOCKS AND 4433 03:37:48,467 --> 03:37:52,037 REACTION TEMPLATES THAT ENCODE 4434 03:37:52,037 --> 03:37:53,338 ROBUST TRANSFORMATIONS SOUNDS 4435 03:37:53,338 --> 03:37:56,341 FAMILIAR, FOR THOSE FAMILIAR 4436 03:37:56,341 --> 03:37:59,244 WITH THE LIBRARIES THIS IS 4437 03:37:59,244 --> 03:38:00,812 EXACTLY HOW WE CONSTRUCT THE 4438 03:38:00,812 --> 03:38:01,613 LIBRARIES. 4439 03:38:01,613 --> 03:38:07,385 WE USE THESE RULES TO ENUMERATE 4440 03:38:07,385 --> 03:38:09,387 CHEMICAL REACTIONS TO DEFINE THE 4441 03:38:09,387 --> 03:38:12,624 SPACE IN MILLIONS OR TRILLIONS 4442 03:38:12,624 --> 03:38:16,828 OR MORE MOLECULAR STRUCTURES 4443 03:38:16,828 --> 03:38:20,132 SYNTHETICALLY ACCESSIBLE. 4444 03:38:20,132 --> 03:38:28,840 THE USE OF DEEP LEARNING IS 4445 03:38:28,840 --> 03:38:32,244 RATHER THAN WORRYING ENUMERATION 4446 03:38:32,244 --> 03:38:33,712 WE'RE TEACHING MODELS TO 4447 03:38:33,712 --> 03:38:34,780 NAVIGATE THE SPACE AND WHATEVER 4448 03:38:34,780 --> 03:38:36,815 GOAL WE'RE TRYING TO OPTIMIZE 4449 03:38:36,815 --> 03:38:38,817 FOR THEY KNOW HOW TO NAVIGATE 4450 03:38:38,817 --> 03:38:41,019 THE LARGE SPACES WITHOUT EVER 4451 03:38:41,019 --> 03:38:42,687 WORRYING HOW MUCH MOLECULES ARE 4452 03:38:42,687 --> 03:38:44,856 IN SIDE THE SPACE AND LEARN HOW 4453 03:38:44,856 --> 03:38:46,725 TO DO THE SELECTION PROCESS AND 4454 03:38:46,725 --> 03:38:49,161 ON THE FLY GENERATE THE 4455 03:38:49,161 --> 03:38:52,731 REACTIONS AND MOLECULES IT 4456 03:38:52,731 --> 03:38:57,269 THINKS IT NICE TO TO FIND AN 4457 03:38:57,269 --> 03:38:57,502 OPTIMUM. 4458 03:38:57,502 --> 03:39:01,139 WE HAVE TAKEN THE IDEA IN THE 4459 03:39:01,139 --> 03:39:06,311 APPROACHES AND APPLIED TO DEEP 4460 03:39:06,311 --> 03:39:14,419 LEARNING APPROACHES. 4461 03:39:14,419 --> 03:39:16,221 WE'RE TAKING IN A GRAPH 4462 03:39:16,221 --> 03:39:17,923 STRUCTURE OF A MOLECULE WE'D 4463 03:39:17,923 --> 03:39:19,624 LIKE TO MAKE AND DECODING A 4464 03:39:19,624 --> 03:39:24,830 PATHWAY THAT WE BELIEVE WE CAN 4465 03:39:24,830 --> 03:39:28,834 USE TO MAKE IT. 4466 03:39:28,834 --> 03:39:31,803 TRE TRANSFORMER MODELS ARE USED 4467 03:39:31,803 --> 03:39:36,074 TO PREDICTING SEQUENCES AND 4468 03:39:36,074 --> 03:39:38,944 LETTERS OR TOKENS CONSTRUCTING A 4469 03:39:38,944 --> 03:39:39,211 SENTENCE. 4470 03:39:39,211 --> 03:39:40,712 WE CAN TAKE OUR SYNTHETIC 4471 03:39:40,712 --> 03:39:41,680 PATHWAYS WHICH IN GENERAL ARE 4472 03:39:41,680 --> 03:39:43,348 TREES BECAUSE YOU CAN HAVE 4473 03:39:43,348 --> 03:39:45,984 BRANCHES FOR CONVERTED 4474 03:39:45,984 --> 03:39:47,285 SYNTHESIZES, ETCETERA AND 4475 03:39:47,285 --> 03:39:51,823 LINEARIZE IT OR REPRESENT IT AS 4476 03:39:51,823 --> 03:39:55,093 THIS SEQUENCE AND NOW WE HAVE A 4477 03:39:55,093 --> 03:39:56,061 MODEL ARCHITECTURE WE CAN TRAIN 4478 03:39:56,061 --> 03:39:58,597 OF LARGE SCALE TO TAKE IN 4479 03:39:58,597 --> 03:40:00,298 MOLECULAR STRUCTURES AS TARGETS 4480 03:40:00,298 --> 03:40:05,170 AND OUTPUT THE PATHWAYS THAT 4481 03:40:05,170 --> 03:40:12,844 APPROXIMATELY OR EXACTLY 4482 03:40:12,844 --> 03:40:14,613 RECONSTRUCT THE OUTPUT AND WE 4483 03:40:14,613 --> 03:40:18,149 CAN TAKE IN DESIGNS FOR OTHER 4484 03:40:18,149 --> 03:40:20,252 GENERATIVE MODELS AND PUT THEM 4485 03:40:20,252 --> 03:40:23,688 THROUGH OUR MODEL TO PROPOSE NEW 4486 03:40:23,688 --> 03:40:24,823 STRUCTURES THAT ARE GUARANTEED 4487 03:40:24,823 --> 03:40:28,827 WITHIN THE LIMITS OF OUR 4488 03:40:28,827 --> 03:40:31,529 TEMPLATES TO BE SYNTHETICALLY 4489 03:40:31,529 --> 03:40:31,830 ACCESSIBLE. 4490 03:40:31,830 --> 03:40:35,000 I'M SHOWING A GENERATIVE DESIGN 4491 03:40:35,000 --> 03:40:40,038 FROM AN UNCONSTRAINED MODEL WITH 4492 03:40:40,038 --> 03:40:42,574 A DOCKING SCORE APPROXIMATING A 4493 03:40:42,574 --> 03:40:44,175 BINDING INTERACTION AND PASS 4494 03:40:44,175 --> 03:40:47,812 THIS MOLECULE THROUGH THE 4495 03:40:47,812 --> 03:40:49,114 PROJECTOR MODEL TO RETURN A 4496 03:40:49,114 --> 03:40:51,616 NOVEL STRUCTURE WHERE IT'S 4497 03:40:51,616 --> 03:40:54,486 ASSOCIATED ALREADY WITH A 4498 03:40:54,486 --> 03:40:55,687 SYNTHETIC PATHWAY IN THIS STEP 4499 03:40:55,687 --> 03:40:58,757 ONE STEP COUPLING AND THIS 4500 03:40:58,757 --> 03:40:59,791 APPROACH HOLDS TRUE FOR OTHER 4501 03:40:59,791 --> 03:41:01,259 PROTEINS AND IDEAS. 4502 03:41:01,259 --> 03:41:03,295 WE CAN TAKE THE RECOMMENDATIONS 4503 03:41:03,295 --> 03:41:06,331 FROM UNCONSTRAINED MODELS AND 4504 03:41:06,331 --> 03:41:08,433 QUICKLY PROPOSE ROUTES TO MAKE 4505 03:41:08,433 --> 03:41:12,437 THEM USING COMMERCIALLY 4506 03:41:12,437 --> 03:41:13,772 AVAILABLE BUILDING BLOCKS. 4507 03:41:13,772 --> 03:41:14,506 SO, AGAIN WE'VE BEEN FOLLOWING 4508 03:41:14,506 --> 03:41:16,474 UP THE LINE OF WORK BECAUSE 4509 03:41:16,474 --> 03:41:19,110 REALLY THE NOTION OF GROUNDING 4510 03:41:19,110 --> 03:41:21,846 THE PREDICTIONS OF DEEP LEARNING 4511 03:41:21,846 --> 03:41:24,683 MOLECULAR DESIGN TOOLS IN 4512 03:41:24,683 --> 03:41:25,350 CHEMICAL SYNTHESIS IS ESSENTIAL 4513 03:41:25,350 --> 03:41:27,152 FOR MINIMIZING THE GAP BETWEEN 4514 03:41:27,152 --> 03:41:28,820 THE WHITE BOARD IDEAS OR IDEAS 4515 03:41:28,820 --> 03:41:32,824 WE CAN DRAW UPON THE SCREEN AND 4516 03:41:32,824 --> 03:41:33,692 THE PHYSICAL WORLD. 4517 03:41:33,692 --> 03:41:34,993 A LOT OF TEXT OR IMAGE 4518 03:41:34,993 --> 03:41:37,395 GENERATION PUBLICATIONS OF 4519 03:41:37,395 --> 03:41:38,730 GENERATIVE DESIGN LIVE AND DIE 4520 03:41:38,730 --> 03:41:40,832 ON THE SCREEN. 4521 03:41:40,832 --> 03:41:43,435 YOU DON'T NEED TO PHYSICALLY 4522 03:41:43,435 --> 03:41:47,639 SUBSTANTIATE AN OBJECT FROM 4523 03:41:47,639 --> 03:41:51,977 CLASS GPT OR STABLE DIFFUSION 4524 03:41:51,977 --> 03:41:55,647 BUT WE HAVE TO PHYSICALLY 4525 03:41:55,647 --> 03:41:57,916 PRODUCE THEM TO GET ANSWERS ON 4526 03:41:57,916 --> 03:41:59,017 THE PROPERTIES. 4527 03:41:59,017 --> 03:42:02,954 THE LATEST MODEL IS CALLED SIP 4528 03:42:02,954 --> 03:42:04,189 FORMER AND CONTINUED THE IDEA OF 4529 03:42:04,189 --> 03:42:06,624 THE ARCHITECTURE WHERE WE TAKE 4530 03:42:06,624 --> 03:42:08,860 IN REFERENCE MOLECULE AS INPUTS 4531 03:42:08,860 --> 03:42:13,765 AND KEY CODE THE PATHWAY IN OUR 4532 03:42:13,765 --> 03:42:19,504 LINEARIZED POST-FIXED NOTATION 4533 03:42:19,504 --> 03:42:21,339 CORRESPONDING 1:1 TO THE GRAPH 4534 03:42:21,339 --> 03:42:25,510 AND HAVE THE DECODER MODEL WHERE 4535 03:42:25,510 --> 03:42:27,112 WE DON'T HAVE TO CONDITION ON A 4536 03:42:27,112 --> 03:42:27,712 SPECIFIC REFERENCE. 4537 03:42:27,712 --> 03:42:28,913 WE CAN START GENERATING NEW 4538 03:42:28,913 --> 03:42:33,184 STRUCTURES FROM SCRATCH. 4539 03:42:33,184 --> 03:42:36,421 THE WAY WE'VE BEEN APPLYING THIS 4540 03:42:36,421 --> 03:42:39,024 IS TO DO THIS ANALOG GENERATION 4541 03:42:39,024 --> 03:42:43,628 TAKING ANY DESIGN STRUCTURE AND 4542 03:42:43,628 --> 03:42:48,500 MAP IT BACK TO SYNTHETICALLY 4543 03:42:48,500 --> 03:42:54,305 ACCESSIBLE CHEMICAL SPACE AND 4544 03:42:54,305 --> 03:42:59,044 THIS BORONS IN STRANGE PLACE AND 4545 03:42:59,044 --> 03:43:00,412 WHAT YOU WOULDN'T EXPECT TO FIND 4546 03:43:00,412 --> 03:43:05,216 HERE AND TAKING FROM A 4547 03:43:05,216 --> 03:43:09,354 GENERATIVE MODEL AND AS YOU PASS 4548 03:43:09,354 --> 03:43:12,057 IT THROUGH IT SANITIZES IT AND 4549 03:43:12,057 --> 03:43:13,258 CLEANS IT UP INTO A NEW 4550 03:43:13,258 --> 03:43:14,959 STRUCTURE WHERE WE KNOW HOW TO 4551 03:43:14,959 --> 03:43:19,464 MAKE THE MOLECULE AND PRESERVED 4552 03:43:19,464 --> 03:43:20,832 THE PREDICTED PROPERTY SCORE. 4553 03:43:20,832 --> 03:43:22,867 SO HOW GOOD IS THE STRUCTURE WE 4554 03:43:22,867 --> 03:43:24,803 PRESERVED THROUGH THE ANALOGING 4555 03:43:24,803 --> 03:43:27,238 APPROACH. 4556 03:43:27,238 --> 03:43:31,709 LIKEWISE WE CAN TAKE THE 4557 03:43:31,709 --> 03:43:32,410 UNSATURATED PRODUCT WITH THE 4558 03:43:32,410 --> 03:43:34,746 TEST RESULT WE KNOW HOW TO MAKE 4559 03:43:34,746 --> 03:43:40,218 AND CORRESPONDS TO I A 4560 03:43:40,218 --> 03:43:42,153 COMMERCIALLY AVAILABLE BUILDING 4561 03:43:42,153 --> 03:43:43,788 BLOCK AND USE DOCKING SCORES AS 4562 03:43:43,788 --> 03:43:46,024 OUR OBJECTIVE FUNCTION TO 4563 03:43:46,024 --> 03:43:46,458 OPTIMIZE. 4564 03:43:46,458 --> 03:43:48,093 SOMETIMES WE'LL GET A SLIGHTLY 4565 03:43:48,093 --> 03:43:49,227 WORSE SCORE, SOMETIMES BETTER 4566 03:43:49,227 --> 03:43:50,795 SCORE BUT THE POINT IS WE CAN 4567 03:43:50,795 --> 03:43:52,664 TAKE THE IDEAS FROM ANY SOURCE 4568 03:43:52,664 --> 03:43:54,566 WE WANT AND CLEAN THEM UP INTO A 4569 03:43:54,566 --> 03:43:59,237 MOLECULE WE KNOW HOW TO MAKE. 4570 03:43:59,237 --> 03:44:05,410 SO THIS NOTION OF SYNTHESIZABLE 4571 03:44:05,410 --> 03:44:09,514 HIT SYNTHESIS CONTINUES IN HIT 4572 03:44:09,514 --> 03:44:11,916 EXPANSION STAGES, HIT TO LEAD 4573 03:44:11,916 --> 03:44:13,151 AND LEAD OPTIMIZATION TO QUICKLY 4574 03:44:13,151 --> 03:44:17,889 TAKE WAY MOLECULE AS A REFERENCE 4575 03:44:17,889 --> 03:44:19,891 POINT AND EXPAND OUT THE 4576 03:44:19,891 --> 03:44:22,427 CHEMICAL SPACE AROUND IT WITH A 4577 03:44:22,427 --> 03:44:27,999 LARGE NUMBER OF SYNTHETICALLY 4578 03:44:27,999 --> 03:44:29,000 TRACTABLE AND SIMILAR 4579 03:44:29,000 --> 03:44:31,936 STRUCTURES. 4580 03:44:31,936 --> 03:44:34,472 SO, SYNTHESIS AGAIN TO EMPHASIZE 4581 03:44:34,472 --> 03:44:36,641 THIS POINT, IT INFLUENCES THE 4582 03:44:36,641 --> 03:44:38,476 KIND OF MOLECULES WE WANT TO 4583 03:44:38,476 --> 03:44:40,078 DESIGN BECAUSE I DON'T WANT TO 4584 03:44:40,078 --> 03:44:41,579 WASTE OUR TIME DESIGNING 4585 03:44:41,579 --> 03:44:42,480 STRUCTURES THAT CAN'T EVER BE 4586 03:44:42,480 --> 03:44:46,084 BROUGHT TO THE LAB AND TESTED. 4587 03:44:46,084 --> 03:44:48,820 SO IN THE VIRTUAL LIBRARY OR 4588 03:44:48,820 --> 03:44:50,788 SCREENING SETTING THAT'S WHY 4589 03:44:50,788 --> 03:44:51,923 WITH WE HAVE THE LIBRARIES 4590 03:44:51,923 --> 03:44:53,057 BECAUSE THEY'RE CONSTRUCTED 4591 03:44:53,057 --> 03:44:54,826 THROUGH REACTIONS WE BELIEVE WE 4592 03:44:54,826 --> 03:44:55,593 CAN RUN. 4593 03:44:55,593 --> 03:44:57,595 WHEN WE LOOKED AT THE GENERATIVE 4594 03:44:57,595 --> 03:44:59,497 MODELS THEY HAVE THE FAILURE 4595 03:44:59,497 --> 03:45:02,200 MODE WHERE IF YOU DON'T TEACH 4596 03:45:02,200 --> 03:45:02,901 GENERATIVE MODELS ABOUT 4597 03:45:02,901 --> 03:45:04,669 SYNTHESIS THEY CAN SOMETIMES GO 4598 03:45:04,669 --> 03:45:06,538 OFF THE RAILS AND PROPOSE 4599 03:45:06,538 --> 03:45:08,606 STRANGE STRUCTURES THAT PASS THE 4600 03:45:08,606 --> 03:45:12,510 MOST BASIC FILTERS OF CHEMICAL 4601 03:45:12,510 --> 03:45:13,144 VALIDITY BUT REALLY DON'T MAKE 4602 03:45:13,144 --> 03:45:19,918 SENSE. 4603 03:45:19,918 --> 03:45:21,853 NOW, SYNTHESIS IS NOT TRULY A 4604 03:45:21,853 --> 03:45:24,622 BINARY PROPERTY OF A MOLECULE OR 4605 03:45:24,622 --> 03:45:29,093 WE CAN VERSUS NOT MAKE THIS IT 4606 03:45:29,093 --> 03:45:31,029 DOESN'T INFLUENCE ENTIRELY THE 4607 03:45:31,029 --> 03:45:32,530 COST OF THE STRUCTURES. 4608 03:45:32,530 --> 03:45:36,701 WE WANT MORE NUANCE WHEN WE 4609 03:45:36,701 --> 03:45:39,304 THINK ABOUT THE SYNTHETIC 4610 03:45:39,304 --> 03:45:40,338 ACCESSIBILITY OF DESIGN 4611 03:45:40,338 --> 03:45:40,805 MOLECULES IN SILICO. 4612 03:45:40,805 --> 03:45:43,808 THAT BRINGS US TO THE FINAL 4613 03:45:43,808 --> 03:45:46,644 VIGNETTE STORY I'D LIKE TO SHARE 4614 03:45:46,644 --> 03:45:48,813 IN THE LAST FEW MINUTES. 4615 03:45:48,813 --> 03:45:52,517 IT'S A TOOL DESIGN TO TAKE IN 4616 03:45:52,517 --> 03:45:54,285 SUCTIONS FROM VIRTUALLY ANY 4617 03:45:54,285 --> 03:45:54,586 SOURCE. 4618 03:45:54,586 --> 03:45:57,755 THESE DAYS WE HAVE OUR VIRTUAL 4619 03:45:57,755 --> 03:45:58,523 LIBRARIES, WE HAVE DE NOVO 4620 03:45:58,523 --> 03:46:00,825 DESIGN TOOLS AND EXPERTS WITH 4621 03:46:00,825 --> 03:46:05,029 IDEAS AND CAN ENUMERATE 4622 03:46:05,029 --> 03:46:06,531 CANDIDATE STRUCTURES. 4623 03:46:06,531 --> 03:46:08,099 IF WE'RE TRYING TO UNDERSTAND 4624 03:46:08,099 --> 03:46:11,069 FROM THE IDEAS WHICH ARE 4625 03:46:11,069 --> 03:46:13,238 VIRTUALLY INFINITE IN NUMBER, 4626 03:46:13,238 --> 03:46:14,939 WE'RE TRYING TO UNDERSTAND WHICH 4627 03:46:14,939 --> 03:46:16,908 ONES SHOULD WE PRIORITIZE FOR 4628 03:46:16,908 --> 03:46:20,144 SYNTHESIS AND EVALUATION. 4629 03:46:20,144 --> 03:46:24,015 WE SHOULD DO IT ENTIRELY ON 4630 03:46:24,015 --> 03:46:25,116 PREDICTIVE PROPERTIES BECAUSE IT 4631 03:46:25,116 --> 03:46:27,118 COULD BE ONE MOLECULE WE CAN 4632 03:46:27,118 --> 03:46:28,386 PURCHASE QUICKLY AND ANOTHER 4633 03:46:28,386 --> 03:46:32,824 MOLECULE THAT WOULD TAKE A Ph.D. 4634 03:46:32,824 --> 03:46:35,393 AND ORGANIC CHEMISTRY LAB AND 4635 03:46:35,393 --> 03:46:36,928 WANT TO UNDERSTAND THIS AS MORE 4636 03:46:36,928 --> 03:46:40,865 OF THE FLUID VALUE AND 4637 03:46:40,865 --> 03:46:42,834 CONTINUOUS VALUE. 4638 03:46:42,834 --> 03:46:44,135 SO THIS FRAMEWORK WE CALL 4639 03:46:44,135 --> 03:46:45,903 SPARROW TRIES TO TAKE ALL THE 4640 03:46:45,903 --> 03:46:53,077 CON -- CANDIDATES AND TO A SET 4641 03:46:53,077 --> 03:46:53,945 THAT STRIKES A BALANCE BETWEEN 4642 03:46:53,945 --> 03:46:56,481 THE REWARD OR BENEFIT OF TESTING 4643 03:46:56,481 --> 03:46:59,050 THE MOLECULES ACCORDING TO 4644 03:46:59,050 --> 03:47:00,918 PROPERTY PREDICTION MODELS AND 4645 03:47:00,918 --> 03:47:03,521 THEIR SNBT COST. 4646 03:47:03,521 --> 03:47:04,656 -- SYNTHETIC COST. 4647 03:47:04,656 --> 03:47:09,694 AND WHAT IS INTERESTING IS WE 4648 03:47:09,694 --> 03:47:12,030 CONSIDER THINGS LIKE SHARED 4649 03:47:12,030 --> 03:47:13,164 INTERMEDIATES AND USING PARALLEL 4650 03:47:13,164 --> 03:47:13,798 CHEMISTRY. 4651 03:47:13,798 --> 03:47:16,434 NOT ALL MOLECULES COST THE SAME. 4652 03:47:16,434 --> 03:47:18,569 NOT ALL REACTION PATHWAYS THAT 4653 03:47:18,569 --> 03:47:20,471 HAVE THE SAME NUMBER OF STEPS 4654 03:47:20,471 --> 03:47:22,774 ARE THE SAME IN PRACTICE. 4655 03:47:22,774 --> 03:47:24,442 SO SPARROW'S TRYING TO MAKE THE 4656 03:47:24,442 --> 03:47:26,678 DECISIONS IN THE TRADE-OFFS IN A 4657 03:47:26,678 --> 03:47:27,211 QUANTITATIVE WAY. 4658 03:47:27,211 --> 03:47:30,648 BECAUSE WE WANTED TO BE A 4659 03:47:30,648 --> 03:47:31,516 QUANTITATIVE OPTIMIZATION IT 4660 03:47:31,516 --> 03:47:34,319 REQUIRES US TO DEFINE VARIABLES 4661 03:47:34,319 --> 03:47:35,887 IN THE OPTIMIZATION. 4662 03:47:35,887 --> 03:47:39,390 WE'RE GOING TO SIMPLIFY THINGS 4663 03:47:39,390 --> 03:47:48,333 AND FOCUS ON THE INTERWIGS -- 4664 03:47:48,333 --> 03:47:50,101 INTUITION IN THE SELECTION 4665 03:47:50,101 --> 03:47:51,836 PROCESS IT'S TRYING TO MAXIMIZE 4666 03:47:51,836 --> 03:47:55,440 THE BENEFIT OR PROPERTIES WE SEE 4667 03:47:55,440 --> 03:47:56,974 IN THE STRUCTURES AND TRYING TO 4668 03:47:56,974 --> 03:47:59,477 MINIMIZE THE COST IN TERMS OF 4669 03:47:59,477 --> 03:48:02,580 BUILDING BLOCK COSTS OR CHEMICAL 4670 03:48:02,580 --> 03:48:04,816 COSTS AND FOR MOLECULES THAT 4671 03:48:04,816 --> 03:48:07,251 REQUIRE SYNTHESIS, WE'RE 4672 03:48:07,251 --> 03:48:08,553 PREDICTING THE LIKELIHOOD OF 4673 03:48:08,553 --> 03:48:09,687 SUCCESS AND PATHWAYS USED AND 4674 03:48:09,687 --> 03:48:12,156 TRYING TO MINIMIZE THE NUMBER OF 4675 03:48:12,156 --> 03:48:13,524 REACTIONS WE'RE RUNNING AND WHEN 4676 03:48:13,524 --> 03:48:15,727 WE HAVE DO RUN REACTIONS TRYING 4677 03:48:15,727 --> 03:48:17,161 TO MAXIMIZE THE LIKELIHOOD OF 4678 03:48:17,161 --> 03:48:19,964 SUCCESS PREDICTED FROM OUR OTHER 4679 03:48:19,964 --> 03:48:22,100 CHEMISTRY TOOLS. 4680 03:48:22,100 --> 03:48:24,268 SO IN THE PAPER WE DESCRIBED 4681 03:48:24,268 --> 03:48:25,536 CASE STUDIES AND FOCUSSING ON 4682 03:48:25,536 --> 03:48:27,972 ONE HERE WE LOOKED AT A DESIGN 4683 03:48:27,972 --> 03:48:29,173 CYCLE WITH 121 STRUCTURES UNDER 4684 03:48:29,173 --> 03:48:32,643 CONSIDERATION. 4685 03:48:32,643 --> 03:48:35,913 THIS IS AN AUTONOMOUS DISCOVERY 4686 03:48:35,913 --> 03:48:38,783 OF DYE MOLECULES AND COULDN'T 4687 03:48:38,783 --> 03:48:46,958 MAKE 121 IN RAY -- A SINGLE 4688 03:48:46,958 --> 03:48:47,225 ITERATION. 4689 03:48:47,225 --> 03:48:52,830 IN THIS EXAMPLE FROM 121 4690 03:48:52,830 --> 03:48:53,464 CANDID 4691 03:48:53,464 --> 03:48:58,469 CANDIDATES, SPARROW SAID I THINK 4692 03:48:58,469 --> 03:49:02,073 YOU SHOULD TEST THESE 15 BASED 4693 03:49:02,073 --> 03:49:06,944 ON YOUR INFORMATION. 4694 03:49:06,944 --> 03:49:10,181 SIX SPARROW DETERMINED COULD BE 4695 03:49:10,181 --> 03:49:11,783 PURCHASED AND TESTED. 4696 03:49:11,783 --> 03:49:13,384 THERE WERE NINE ADDITIONAL 4697 03:49:13,384 --> 03:49:15,720 MOLECULE FROM THE LARGE SET 4698 03:49:15,720 --> 03:49:19,957 SPARROW SUGGESTED WE MAKE. 4699 03:49:19,957 --> 03:49:22,160 MAKING THE NINE REQUIRED SIX 4700 03:49:22,160 --> 03:49:23,628 BUILDING BLOCKS AND 10 REACTION 4701 03:49:23,628 --> 03:49:23,895 STEPS. 4702 03:49:23,895 --> 03:49:26,230 IF YOU LOOK AT THE PATHWAYS AND 4703 03:49:26,230 --> 03:49:27,899 PINK STRUCTURES, YOU MIGHT 4704 03:49:27,899 --> 03:49:29,834 RECOGNIZE THAT WE'RE USING 4705 03:49:29,834 --> 03:49:30,601 SIMILAR STARTING MATERIALS FROM 4706 03:49:30,601 --> 03:49:33,504 MULTIPLE CANDIDATES. 4707 03:49:33,504 --> 03:49:34,939 WE'RE TESTING INTERMEDIATE ON 4708 03:49:34,939 --> 03:49:39,143 THE WAY TO FINAL PRODUCT AND 4709 03:49:39,143 --> 03:49:39,811 GETTING COLLOQUIALLY BANG FOR 4710 03:49:39,811 --> 03:49:43,314 OUR BUCK AND TESTING A LOT OF 4711 03:49:43,314 --> 03:49:45,616 STRUCTURES WITH A SMALL NUMBER 4712 03:49:45,616 --> 03:49:47,418 OF SYNTHETIC EFFORT WHILE ABLE 4713 03:49:47,418 --> 03:49:49,487 TO ACCESS THIS INTERESTING SET 4714 03:49:49,487 --> 03:49:52,824 OF 15 CANDIDATES IN THE DESIGN 4715 03:49:52,824 --> 03:49:54,992 CYCLE. 4716 03:49:54,992 --> 03:49:56,360 SUMMARIZING THE STORIES WE THINK 4717 03:49:56,360 --> 03:50:00,131 ABOUT DIFFERENT WAYS OF 4718 03:50:00,131 --> 03:50:01,966 PROPOSING NEW MOLECULAR 4719 03:50:01,966 --> 03:50:03,234 STRUCTURES TO DRIVE DESIGN 4720 03:50:03,234 --> 03:50:05,069 CYCLES AND IN DO SO WE DON'T 4721 03:50:05,069 --> 03:50:08,005 WANT TO END UP WITH 4722 03:50:08,005 --> 03:50:15,646 UNSYNTHESIZABLE OR INACCESSIBLE 4723 03:50:15,646 --> 03:50:17,415 STRUCTURES AND USING TOOLS TO 4724 03:50:17,415 --> 03:50:19,183 PREDICT THE SUCCESSION OF 4725 03:50:19,183 --> 03:50:20,818 EXPERIMENTAL VALIDATION WE THINK 4726 03:50:20,818 --> 03:50:23,721 OF BIASSING OR CONSTRAINING 4727 03:50:23,721 --> 03:50:25,823 GENERATIVE A.I. TO HAVE TO THINK 4728 03:50:25,823 --> 03:50:27,525 ABOUT SYNTHETIC PATHWAYS NOT 4729 03:50:27,525 --> 03:50:28,826 JUST THINK ABOUT STRUCTURES. 4730 03:50:28,826 --> 03:50:31,929 AND WE ALSO THINK ABOUT THE 4731 03:50:31,929 --> 03:50:35,099 OPTIMAL DESIGN FRAMEWORKS THAT 4732 03:50:35,099 --> 03:50:40,805 LET US BALANCE INFORMATION GAIN 4733 03:50:40,805 --> 03:50:42,273 VERSUS STOCK 4734 03:50:42,273 --> 03:50:44,509 VER 4735 03:50:44,509 --> 03:50:44,976 VER 4736 03:50:44,976 --> 03:50:47,078 VERSUS SYNTHETIC COSTS. 4737 03:50:47,078 --> 03:50:49,280 I'LL ACKNOWLEDGE ALL THE 4738 03:50:49,280 --> 03:50:52,250 FANTASTIC CONTRIBUTIONS FROM THE 4739 03:50:52,250 --> 03:50:54,151 GROUP THAT MADE THIS POSSIBLE 4740 03:50:54,151 --> 03:50:56,821 AND GENEROUS FUNDING SUPPORT AND 4741 03:50:56,821 --> 03:50:58,122 FOR QUESTIONS I'LL FLIP BACK SO 4742 03:50:58,122 --> 03:51:04,028 THANK YOU FOR THE INVITATION AND 4743 03:51:04,028 --> 03:51:08,833 TIME. 4744 03:51:08,833 --> 03:51:12,803 >> THANK YOU FOR AN EXTREMELY 4745 03:51:12,803 --> 03:51:14,906 CLEAR TALK. 4746 03:51:14,906 --> 03:51:17,942 LET ME START WITH COMMENTS TO 4747 03:51:17,942 --> 03:51:18,376 SHARE. 4748 03:51:18,376 --> 03:51:22,847 THANK YOU FOR SHARING GENEROUSLY 4749 03:51:22,847 --> 03:51:24,815 THE TOOLS YOU DEVELOPED. 4750 03:51:24,815 --> 03:51:28,386 SOMEONE WAS SAYING IN THE Q&A 4751 03:51:28,386 --> 03:51:29,420 IT'S EXTREMELY EDUCATIONAL FROM 4752 03:51:29,420 --> 03:51:29,687 THE FIELD. 4753 03:51:29,687 --> 03:51:40,064 THANK YOU FOR SHARING. 4754 03:51:47,939 --> 03:51:52,810 A FEW QUESTIONS. 4755 03:51:52,810 --> 03:51:58,115 >> SO THE CONNECTION BETWEEN 4756 03:51:58,115 --> 03:52:02,119 WHAT WE'RE DOING AND THE WAY 4757 03:52:02,119 --> 03:52:04,255 THEY'RE CONSTRUCTED. 4758 03:52:04,255 --> 03:52:07,825 THE PHILOSOPHY AND CONSTRAINTS 4759 03:52:07,825 --> 03:52:08,225 ARE IDENTICAL. 4760 03:52:08,225 --> 03:52:10,494 THE IDEA IS AS THE FIELD 4761 03:52:10,494 --> 03:52:11,429 CONVERGES ON UNDERSTANDING OF 4762 03:52:11,429 --> 03:52:13,397 WHAT TRANSFORMATIONS MAY BE 4763 03:52:13,397 --> 03:52:14,732 POSSIBLE AND WHAT BUILDING 4764 03:52:14,732 --> 03:52:16,734 BLOCKS ARE COLLECTIVELY 4765 03:52:16,734 --> 03:52:19,604 AVAILABLE, OUR INTENTION IS NOT 4766 03:52:19,604 --> 03:52:23,941 TO NECESSARILY REDEFINE THE 4767 03:52:23,941 --> 03:52:25,276 IDEAS AND MAKE DIFFERENT 4768 03:52:25,276 --> 03:52:29,680 DECISIONS BUT CHANGE HOW WE'RE 4769 03:52:29,680 --> 03:52:31,649 NAVIGATING THE SPACES. 4770 03:52:31,649 --> 03:52:34,585 AND I'VE ARGUED WITH DEEP 4771 03:52:34,585 --> 03:52:35,553 GENERATIVE APPROACHES ONE 4772 03:52:35,553 --> 03:52:38,956 DOESN'T HAVE TO DO ENUMERATION 4773 03:52:38,956 --> 03:52:40,758 TO BENEFIT FROM THE SPACE THEY 4774 03:52:40,758 --> 03:52:40,992 DESIGN. 4775 03:52:40,992 --> 03:52:42,960 AS THEY GET BIGGER AND BIGGER 4776 03:52:42,960 --> 03:52:46,931 WE'RE NOT GOING TO WANT TO STORE 4777 03:52:46,931 --> 03:52:57,508 THEM IN A NUMERATIVE FASHION. 4778 03:53:05,583 --> 03:53:08,185 >> HOW YOU NAVIGATE. 4779 03:53:08,185 --> 03:53:10,388 SO MORE ON THE PRACTICAL SIDE. 4780 03:53:10,388 --> 03:53:12,823 MOST YOUR TALK HAS BEEN 4781 03:53:12,823 --> 03:53:14,125 METHODOLOGICAL. 4782 03:53:14,125 --> 03:53:17,061 DO YOU HAVE SUCCESS WITH DESIGN 4783 03:53:17,061 --> 03:53:22,099 AND MORE ACTIVE COMPOUNDS THAT 4784 03:53:22,099 --> 03:53:25,469 MAY BE OUTSIDE OF THE 4785 03:53:25,469 --> 03:53:25,803 PRESENTATION. 4786 03:53:25,803 --> 03:53:28,906 >> WE HAVE SEVERAL ONGOING 4787 03:53:28,906 --> 03:53:31,308 COLLABORATIONS AND VALIDATION 4788 03:53:31,308 --> 03:53:33,778 CAMPAIGNS TO SEE HOW THESE 4789 03:53:33,778 --> 03:53:34,812 VALIDATES. 4790 03:53:34,812 --> 03:53:36,414 I'D SAY THERE'S A CONVENIENCE 4791 03:53:36,414 --> 03:53:38,616 FACTOR CERTAINLY IN DOING THE 4792 03:53:38,616 --> 03:53:40,818 INITIAL METHOD DEVELOPMENT WITH 4793 03:53:40,818 --> 03:53:41,952 COMPUTATIONAL EVALUATION BUT 4794 03:53:41,952 --> 03:53:44,655 THAT'S THE INTEREST MOVING 4795 03:53:44,655 --> 03:53:44,889 FORWARD. 4796 03:53:44,889 --> 03:53:47,324 WE HAVE COLLABORATIONS TO TALK 4797 03:53:47,324 --> 03:53:49,260 ABOUT BUT THAT'S THE HOPE. 4798 03:53:49,260 --> 03:53:51,228 AND PERHAPS IF I GIVE A SIMILAR 4799 03:53:51,228 --> 03:53:52,797 TALK IN A COUPLE YEARS YOU'LL 4800 03:53:52,797 --> 03:53:58,402 SEE A FEW MORE IC50 VALUES ON MY 4801 03:53:58,402 --> 03:53:59,837 SLIDE AND WILL HAVE MORE SUCCESS 4802 03:53:59,837 --> 03:54:00,805 VALUES TO SHARE. 4803 03:54:00,805 --> 03:54:01,672 >> A COUPLE QUICK QUESTIONS FROM 4804 03:54:01,672 --> 03:54:08,679 ME. 4805 03:54:08,679 --> 03:54:13,818 ANY COMMENTARY ON PREDICTIVE 4806 03:54:13,818 --> 03:54:14,018 YIELD? 4807 03:54:14,018 --> 03:54:15,219 THAT'S SOMETHING WE RARELY TALK 4808 03:54:15,219 --> 03:54:15,419 ABOUT. 4809 03:54:15,419 --> 03:54:20,825 THAT'S THE FIRST ONE. 4810 03:54:20,825 --> 03:54:22,259 A LOT OF THOUGHTS ABOUT 4811 03:54:22,259 --> 03:54:22,793 PREDICTING YIELD. 4812 03:54:22,793 --> 03:54:24,428 FEW WILL FIT IN THE NEXT 30 4813 03:54:24,428 --> 03:54:25,429 SECONDS OR ONE MINUTE. 4814 03:54:25,429 --> 03:54:30,634 THERE'S CONFOUNDING VARIABLES 4815 03:54:30,634 --> 03:54:37,007 WHEN YOU LOOK AT LITERATURE DATA 4816 03:54:37,007 --> 03:54:42,947 SETS AND HOLD CONCENTRATIONS. 4817 03:54:42,947 --> 03:54:46,217 YOU CAN HAVE COURSE MODELS OR 4818 03:54:46,217 --> 03:54:48,552 HIGH PERFORMING MODELS WITH 4819 03:54:48,552 --> 03:54:49,320 NARROW APPLICABILITY AND IT'S 4820 03:54:49,320 --> 03:54:52,823 DIFFICULT TO BRIDGE THE TWO AT 4821 03:54:52,823 --> 03:55:03,033 THE MOMENT. 4822 03:55:08,739 --> 03:55:11,075 >> WE HAVE JOHN CHODERA. 4823 03:55:11,075 --> 03:55:15,813 HE'S A MEMBER OF THE 4824 03:55:15,813 --> 03:55:22,720 COMPUTATIONAL PROGRAM AT 4825 03:55:22,720 --> 03:55:24,522 SLOAN-KETTERING CANCER CENTER IN 4826 03:55:24,522 --> 03:55:26,991 NEW YORK CITY AND LOOKING AT 4827 03:55:26,991 --> 03:55:30,327 DRUG DISCOVERY POUFRD BY THE 4828 03:55:30,327 --> 03:55:32,163 NEXT GENERATION HYBRID MODELS 4829 03:55:32,163 --> 03:55:37,468 AND USED TO PREDICT RESISTANCE 4830 03:55:37,468 --> 03:55:39,470 IN A DATA MANNER SOMETHING I 4831 03:55:39,470 --> 03:55:41,071 WANT TO ADMIRE AND HIGHLIGHT IS 4832 03:55:41,071 --> 03:55:44,308 HE'S BEEN ONE OF THE BIGGEST 4833 03:55:44,308 --> 03:55:47,378 COMPONENTS AND CONTRIBUTORS TO 4834 03:55:47,378 --> 03:55:57,922 OPEN SCIENCE DEVELOPING OPEN NM 4835 03:56:04,895 --> 03:56:11,068 AND LOOKED AT IS THE MOON SHOT 4836 03:56:11,068 --> 03:56:14,004 AND THE ANTIVIRAL PROGRAMS. 4837 03:56:14,004 --> 03:56:15,206 HE'S AN INVESTIGATOR IN OTHER 4838 03:56:15,206 --> 03:56:21,745 WORKS AND LOOK FORWARD TO THE 4839 03:56:21,745 --> 03:56:21,946 TALK. 4840 03:56:21,946 --> 03:56:24,515 >> THANKS FOR THE INVITATION. 4841 03:56:24,515 --> 03:56:26,217 I'M HOPING TO GIVE PRACTICAL 4842 03:56:26,217 --> 03:56:28,185 ADVICE THAT MAY BE A GOOD 4843 03:56:28,185 --> 03:56:30,454 FOUNDATION FOR YOUR EXPLORATIONS 4844 03:56:30,454 --> 03:56:32,356 IN THE WORKSHOP AS YOU GO OFF 4845 03:56:32,356 --> 03:56:39,463 AND DO MORE COOL STUFF WITH A.I. 4846 03:56:39,463 --> 03:56:44,401 AND ML. 4847 03:56:44,401 --> 03:56:46,604 JUST TO START, I'D LIKE TO 4848 03:56:46,604 --> 03:56:48,472 HIGHLIGHT THE CHALLENGE. 4849 03:56:48,472 --> 03:56:50,641 IT'S SLOW AND COSTLY AND THERE'S 4850 03:56:50,641 --> 03:56:53,077 AN EXTREMELY HIGH FAILURE RATE. 4851 03:56:53,077 --> 03:56:54,678 I TRIED TO LOOK AT THE 4852 03:56:54,678 --> 03:56:57,548 STATISTICS AND YOU CAN FIND A 4853 03:56:57,548 --> 03:56:59,650 COPY HERE AND FOLLOW ALONG. 4854 03:56:59,650 --> 03:57:01,919 I TRIED TO SYNTHESIZE THE 4855 03:57:01,919 --> 03:57:04,822 SUCCESS RATES AND COSTS NOT 4856 03:57:04,822 --> 03:57:06,624 INCLUDING FAILURES AT EACH STAGE 4857 03:57:06,624 --> 03:57:09,360 AND TALK ABOUT HOW TO USE THE 4858 03:57:09,360 --> 03:57:11,595 MODELS TO TEST IN THE DISCOVERY 4859 03:57:11,595 --> 03:57:13,864 PART BUT THE FAILURE RATES AND 4860 03:57:13,864 --> 03:57:15,733 COSTS ARE HIGH FOR THE STAGES. 4861 03:57:15,733 --> 03:57:17,568 SO ANYTHING WE CAN DO IN 4862 03:57:17,568 --> 03:57:20,604 DISCOVERY TO HAVE MORE SHOTS ON 4863 03:57:20,604 --> 03:57:23,274 GOAL OR HIGHER RATES WILL HELP 4864 03:57:23,274 --> 03:57:27,978 IMPROVE THE STATE OF DRUG 4865 03:57:27,978 --> 03:57:28,379 DISCOVERY. 4866 03:57:28,379 --> 03:57:30,714 I SPEAK FROM A LITTLE BIT MUCH 4867 03:57:30,714 --> 03:57:31,015 EXPERIENCE. 4868 03:57:31,015 --> 03:57:36,820 I COLLABORATE WITH COMPANIES AND 4869 03:57:36,820 --> 03:57:43,093 INDUSTRY AND I HAD TO SUPPORT 4870 03:57:43,093 --> 03:57:46,196 THE PROGRAM AS PART OF THIS OPEN 4871 03:57:46,196 --> 03:57:48,799 SCIENCE EXAMPLE CALLED THE COVID 4872 03:57:48,799 --> 03:57:51,201 MOON SHOT UNIQUE AMONG DISCOVERY 4873 03:57:51,201 --> 03:57:52,603 PROJECTS PUTTING ALL THE DATA 4874 03:57:52,603 --> 03:57:52,836 ONLINE. 4875 03:57:52,836 --> 03:57:55,973 YOU CAN GO TO THIS WEBSITE AND 4876 03:57:55,973 --> 03:57:57,441 FIND ALL THE DATA IN A DRUG 4877 03:57:57,441 --> 03:57:58,842 DISCOVERY PROGRAM AND WANT TO 4878 03:57:58,842 --> 03:58:01,345 LOOK AT THE NITTY-GRITTY OF WHAT 4879 03:58:01,345 --> 03:58:03,280 HAPPENS IN THE PROGRAMS I'D 4880 03:58:03,280 --> 03:58:08,819 ENCOURAGE YOU TO BUT WE STARTED 4881 03:58:08,819 --> 03:58:13,257 FROM AN X-RAY FRAGMENT SCREEN 4882 03:58:13,257 --> 03:58:16,827 AND THROUGH A SERIES OF RAPID 4883 03:58:16,827 --> 03:58:21,465 HIT TO LEAD ASSISTED BY TEAMS AS 4884 03:58:21,465 --> 03:58:27,838 WELL AS COMPUTATION WE ABLE TO 4885 03:58:27,838 --> 03:58:30,808 LOOK AT OTHER ABILITIES THAN 4886 03:58:30,808 --> 03:58:34,979 OTHER DRUGS YOU PLAY BE FAMILIAR 4887 03:58:34,979 --> 03:58:36,780 WITH AND MAY HAVE TAKEN. 4888 03:58:36,780 --> 03:58:38,282 SO I'LL USE THIS AS AN EXAMPLE 4889 03:58:38,282 --> 03:58:41,151 AS I TALK ABOUT THE CHALLENGES 4890 03:58:41,151 --> 03:58:42,586 WE'RE FACING. 4891 03:58:42,586 --> 03:58:43,921 IN ANY DRUG DISCOVERY PROGRAM 4892 03:58:43,921 --> 03:58:45,956 YOU HAVE TO HAVE A DESIGN 4893 03:58:45,956 --> 03:58:47,558 OBJECTIVE AND FOR SMALL MOLECULE 4894 03:58:47,558 --> 03:58:50,260 THE TARGET CANDIDATE PROFILE 4895 03:58:50,260 --> 03:58:51,261 THAT SPECIFIES WHAT WE'RE TRYING 4896 03:58:51,261 --> 03:58:52,563 TO ACHIEVE. 4897 03:58:52,563 --> 03:58:53,897 YOU MIGHT WANT TO HIT A 4898 03:58:53,897 --> 03:58:56,734 PARTICULAR TARGET WITH THE SARS 4899 03:58:56,734 --> 03:59:00,804 COV2 PROTEASE AND HAVE TO HIT A 4900 03:59:00,804 --> 03:59:08,278 BETTER AFFINITY THAN 10 NANO 4901 03:59:08,278 --> 03:59:10,114 MOLAR ACTIVITY AND THERE'S OTHER 4902 03:59:10,114 --> 03:59:12,082 PROPERTIES THERE TO ADMINISTER 4903 03:59:12,082 --> 03:59:17,354 IT RATHER THAN OTHER ROUTES IT 4904 03:59:17,354 --> 03:59:21,992 NEEDS TO BE SOLUBLE AND HALF 4905 03:59:21,992 --> 03:59:22,760 LIFE OTHERWISE IT BECOMES 4906 03:59:22,760 --> 03:59:23,961 DIFFICULT TO MANAGE. 4907 03:59:23,961 --> 03:59:28,532 THERE'S A HUGE PANEL OF TOXICITY 4908 03:59:28,532 --> 03:59:29,867 CONCERNS YOU HEARD OF ALREADY 4909 03:59:29,867 --> 03:59:31,568 MAKING IT SAFE ENOUGH TO TAKE 4910 03:59:31,568 --> 03:59:35,105 FOR FIVE OR SEVER EN DAYS. 4911 03:59:35,105 --> 03:59:36,507 -- SEVEN DAYS. 4912 03:59:36,507 --> 03:59:39,109 WE CAN'T RELY ON GENERATIVE 4913 03:59:39,109 --> 03:59:44,815 MODELS THOUGH THERE'S EXCITEMENT 4914 03:59:44,815 --> 03:59:47,584 HOW THEY CAN BE USEFUL. 4915 03:59:47,584 --> 03:59:50,287 WE LOOKED AT THE NUMBER 4916 03:59:50,287 --> 03:59:51,121 ENORMOUS, 21 MILLION 4917 03:59:51,121 --> 03:59:52,256 MEASUREMENTS COMPARED TO THE 4918 03:59:52,256 --> 03:59:54,458 NUMBER OF STRUCTURES IN THE 4919 03:59:54,458 --> 03:59:56,727 PROTEIN DATA BANK WHICH IS 4920 03:59:56,727 --> 03:59:58,796 220,000 AND SIX ORDERS OF 4921 03:59:58,796 --> 04:00:02,766 MAGNITUDES FROM THE TOKENS OF 4922 04:00:02,766 --> 04:00:06,236 TRAINING GPT 4 AND PRODUCING 4923 04:00:06,236 --> 04:00:08,806 USEFUL TEXT IN SUBDOMAINS. 4924 04:00:08,806 --> 04:00:12,076 TO MAKE MORE DATA IN DRUG 4925 04:00:12,076 --> 04:00:14,278 DISCOVERY USING TRADITIONAL 4926 04:00:14,278 --> 04:00:18,449 METHODS WOULD COST AROUND $6 4927 04:00:18,449 --> 04:00:24,254 TRILLION AND UNLIKELY WE'LL GET 4928 04:00:24,254 --> 04:00:25,789 TO THAT SCALE SO INTEREST TO BE 4929 04:00:25,789 --> 04:00:27,391 MORE CLEVER. 4930 04:00:27,391 --> 04:00:28,826 WE ORGANIZED INTO AN ASSAY 4931 04:00:28,826 --> 04:00:29,093 CASCADE. 4932 04:00:29,093 --> 04:00:32,029 WE HAVE ALL THE CHEAP AND EASY 4933 04:00:32,029 --> 04:00:33,897 ASSAYS THAT COME RIGHT AFTER 4934 04:00:33,897 --> 04:00:35,032 SYNTHESIS IN THE FIRST TIER. 4935 04:00:35,032 --> 04:00:36,800 WE STILL HAVE TO MAKE THE 4936 04:00:36,800 --> 04:00:38,035 COMPOUND WHICH IS USUALLY 4937 04:00:38,035 --> 04:00:38,302 DIFFICULT. 4938 04:00:38,302 --> 04:00:42,473 IT COSTS MONEY AND TIME AND THIN 4939 04:00:42,473 --> 04:00:44,174 HAVE TO DO LOGISTICS AND IF IT 4940 04:00:44,174 --> 04:00:48,812 DOES WELL WE PUT IT THROUGH MORE 4941 04:00:48,812 --> 04:00:52,349 EXTENSIVE ASSAYS THAT TELL US 4942 04:00:52,349 --> 04:00:54,751 ABOUT THE PHARMACO KINETICS AND 4943 04:00:54,751 --> 04:00:56,820 IT'S SOMETHING WE WANT TO DO 4944 04:00:56,820 --> 04:00:58,222 WHEN WE HAVE TO. 4945 04:00:58,222 --> 04:01:00,324 SO YOU CAN FIND MORE EXAMPLES OF 4946 04:01:00,324 --> 04:01:04,428 THESE ASSAY CASCADES FROM OUR 4947 04:01:04,428 --> 04:01:08,332 OPEN SCIENCE ANTIVIRAL DRUG 4948 04:01:08,332 --> 04:01:09,433 DISCOVERY PROGRAM AND YOU CAN 4949 04:01:09,433 --> 04:01:12,069 SEE ASSAY CASCADES FOR ACTIVE 4950 04:01:12,069 --> 04:01:13,804 PROGRAMS IF YOU WANT TO EXPLORE 4951 04:01:13,804 --> 04:01:15,272 MORE WHAT THE PROGRAMS LOOK LIKE 4952 04:01:15,272 --> 04:01:18,408 FOR A DIVERSE SET OF PROGRAMS. 4953 04:01:18,408 --> 04:01:20,410 A LOT OF THE TIME YOU'RE 4954 04:01:20,410 --> 04:01:22,946 SPENDING TIME THROUGH THE CYCLES 4955 04:01:22,946 --> 04:01:25,516 TRYING TO MAKE DESIGN IDEAS, 4956 04:01:25,516 --> 04:01:29,019 PRIORITIZING WHICH ONES WILL BE 4957 04:01:29,019 --> 04:01:32,656 SUCCESSFUL AND CUEING THEM FOR 4958 04:01:32,656 --> 04:01:37,194 SYNTHESIS AND GEKT DOWN TO THE 4959 04:01:37,194 --> 04:01:38,996 THINGS TO PUT INTO A PROGRAM TO 4960 04:01:38,996 --> 04:01:39,897 CONVINCE A REGULATORY AGENCY 4961 04:01:39,897 --> 04:01:42,566 LIKE THE FDA YOU HAVE SOMETHING 4962 04:01:42,566 --> 04:01:43,901 SAFE AND EFFICACIOUS ENOUGH TO 4963 04:01:43,901 --> 04:01:45,502 PUT INTO HUMANS. 4964 04:01:45,502 --> 04:01:46,837 THERE'S A LOT OF OPPORTUNITIES 4965 04:01:46,837 --> 04:01:50,807 FOR MODELLING TO IMPACT ALL THE 4966 04:01:50,807 --> 04:01:51,141 STAGES. 4967 04:01:51,141 --> 04:01:55,245 WE'D LIKE TO HAVE MODELS PRE 4968 04:01:55,245 --> 04:01:59,483 PREDICTIVE AND LOOKING AT LIGAND 4969 04:01:59,483 --> 04:02:10,027 PROPERTIES AND SOLUBLE ABILITIES 4970 04:02:11,094 --> 04:02:12,429 AND LOOKING AT THE TARGET AND WE 4971 04:02:12,429 --> 04:02:15,632 PLAY HAVE TO RELY ON SOMETHING 4972 04:02:15,632 --> 04:02:20,404 LIKE A PHYSICAL MODEL AND 4973 04:02:20,404 --> 04:02:22,573 REQUIRES LIGAND BASED TARGET 4974 04:02:22,573 --> 04:02:24,174 WHICH WE PLAY NOT HAVE. 4975 04:02:24,174 --> 04:02:27,010 AND THE INDUSTRY STANDARD WAY TO 4976 04:02:27,010 --> 04:02:28,879 ADVANCE DRUG DISCOVERY BY MAKING 4977 04:02:28,879 --> 04:02:31,982 CHOICES ON WHICH MOLECULES TO 4978 04:02:31,982 --> 04:02:32,282 MAKE. 4979 04:02:32,282 --> 04:02:35,886 THEY'RE ENABLED BY USING 4980 04:02:35,886 --> 04:02:40,057 STRUCTURAL INFORMATION AS SAID 4981 04:02:40,057 --> 04:02:44,861 EARLIER TODAY THAT TELL US HOW 4982 04:02:44,861 --> 04:02:47,798 THEY MAY INTERACT AND IT'S MORE 4983 04:02:47,798 --> 04:02:49,766 THAN DOUBLED IN SIZE FROM 4984 04:02:49,766 --> 04:02:51,101 DEPOSITIONS OF STRUCTURES IN 4985 04:02:51,101 --> 04:02:53,103 PURSUIT OF OTHER SCIENCE GOALS. 4986 04:02:53,103 --> 04:02:55,372 WE'VE SEEN OVER 1,000 NEW 4987 04:02:55,372 --> 04:02:59,109 STRUCTURES IN THE LAST YEARS 4988 04:02:59,109 --> 04:03:02,145 WHICH REMARKABLE AND THERE'S 4989 04:03:02,145 --> 04:03:04,448 IMPORTANT GAPS THAT REMAIN IN 4990 04:03:04,448 --> 04:03:05,616 DIFFERENT THINGS LIKE RELATED 4991 04:03:05,616 --> 04:03:08,018 PROTEINS TO UNDERSTAND WHAT THEY 4992 04:03:08,018 --> 04:03:09,519 LOOK LIKE, FOR EXAMPLE. 4993 04:03:09,519 --> 04:03:11,088 GIVEN THINGS LINING ALPHA FOLD 4994 04:03:11,088 --> 04:03:13,523 WE'VE BEEN ABLE TO DENSELY MODEL 4995 04:03:13,523 --> 04:03:14,758 A LOT OF INTERESTING TARGETS OF 4996 04:03:14,758 --> 04:03:15,058 INTEREST. 4997 04:03:15,058 --> 04:03:17,561 IT'S BECOMING MORE AND MORE OF 4998 04:03:17,561 --> 04:03:19,663 OF A USEFUL RESOURCE WHERE TO 4999 04:03:19,663 --> 04:03:21,031 USE THE MODELLING TECHNIQUES TO 5000 04:03:21,031 --> 04:03:23,867 ANSWER QUESTIONS. 5001 04:03:23,867 --> 04:03:26,103 WHAT THE DESIGN AND CYCLES HAVE 5002 04:03:26,103 --> 04:03:28,739 IS TRYING TO IMPROVE POTENCY. 5003 04:03:28,739 --> 04:03:30,574 THIS IS A WEAK MOLECULE WE 5004 04:03:30,574 --> 04:03:33,477 STARTED WITH OR MAYBE A HIT IN 5005 04:03:33,477 --> 04:03:35,112 THE COVID MOONSHOT AND TRYING TO 5006 04:03:35,112 --> 04:03:39,983 IMPROVE THE POTENCY OR LATER AT 5007 04:03:39,983 --> 04:03:41,151 OTHER ISSUES LIKE METABOLISM 5008 04:03:41,151 --> 04:03:43,754 TRYING TO MAINTAIN POTENCY 5009 04:03:43,754 --> 04:03:44,955 WORKING OUT CHANGES TO CHANGE 5010 04:03:44,955 --> 04:03:46,456 THE PROPERTIES. 5011 04:03:46,456 --> 04:03:48,558 YOU MIGHT START WITH THIS AND 5012 04:03:48,558 --> 04:03:50,560 IDENTIFIED THE ROUTE WHERE 5013 04:03:50,560 --> 04:03:53,530 CHEMISTS HAVE TO OPTIMIZE THE 5014 04:03:53,530 --> 04:03:54,898 ROUTE WHICH SOMETIMES TAKES 5015 04:03:54,898 --> 04:03:56,433 WEEKS AND MAY NOT BE POSSIBLE 5016 04:03:56,433 --> 04:03:58,602 AND MAY COST A LOT OF FOR EACH 5017 04:03:58,602 --> 04:03:59,603 ONE YOU WANT TO SYNTHESIZE AND 5018 04:03:59,603 --> 04:04:04,808 WHAT YOU WANT TO DO IS USE THE 5019 04:04:04,808 --> 04:04:05,942 DIFFERENT STARTING MATERIALS IN 5020 04:04:05,942 --> 04:04:08,178 THE LAST ONE CONNOR SUGGESTED 5021 04:04:08,178 --> 04:04:13,617 HAVING A CHEMIST MAKE A DOZEN 5022 04:04:13,617 --> 04:04:16,086 ANALOGS PER WEEK AND SOME WILL 5023 04:04:16,086 --> 04:04:18,155 MAKE JUST SMALL CHANGES AS 5024 04:04:18,155 --> 04:04:20,390 YOU'LL SEE HERE UP TO ONE PART 5025 04:04:20,390 --> 04:04:24,828 OF THE MOLECULE TO FEEL OUT THE 5026 04:04:24,828 --> 04:04:30,434 STRUCTURE ACTIVITY RELATIONSHIP. 5027 04:04:30,434 --> 04:04:32,803 YOU PLAY HAVE MANY COMPOUNDS 5028 04:04:32,803 --> 04:04:33,570 COMPATIBLE WITH THE LAST 5029 04:04:33,570 --> 04:04:35,105 REACTION THE COUPLING. 5030 04:04:35,105 --> 04:04:35,906 THERE WERE 15,000 POTENTIAL 5031 04:04:35,906 --> 04:04:37,708 GROUPS THAT COULD BE INSTALLED 5032 04:04:37,708 --> 04:04:39,509 AT THE R3 SITE. 5033 04:04:39,509 --> 04:04:40,844 YOU CAN MODEL ALL AND WHAT THEY 5034 04:04:40,844 --> 04:04:45,449 LOOK LIKE IN THE BINDING SITE IN 5035 04:04:45,449 --> 04:04:47,084 A STATIC STRUCTURE BUT WOULD 5036 04:04:47,084 --> 04:04:49,152 LIKE TO PREDICT IN AN ACCURATE 5037 04:04:49,152 --> 04:04:51,254 WAY TO SELECT WHICH COMPOUNDS 5038 04:04:51,254 --> 04:04:55,292 ARE LIKELY TO BE MOST SUCCESSFUL 5039 04:04:55,292 --> 04:04:57,561 AT MAINTAINING POTENCY AND YOU 5040 04:04:57,561 --> 04:04:58,795 MAY NOT HAVE A LOT OF DATA FOR 5041 04:04:58,795 --> 04:05:01,898 THIS PARTICULAR TARGET. 5042 04:05:01,898 --> 04:05:04,801 ONE OF THE BEST PHYSICAL 5043 04:05:04,801 --> 04:05:07,137 MODELLING APPROACHES IS A 5044 04:05:07,137 --> 04:05:07,437 CALCULATION. 5045 04:05:07,437 --> 04:05:09,339 I'LL TELL YOU ABOUT IT BUT 5046 04:05:09,339 --> 04:05:11,074 YOU'RE TRYING TO ASK WHAT IS THE 5047 04:05:11,074 --> 04:05:14,544 DIFFERENCE IN BINDING AFFINITY 5048 04:05:14,544 --> 04:05:18,248 BETWEEN LIGAND A AND B WHERE 5049 04:05:18,248 --> 04:05:23,820 SOME PERTURBATION TO THE 5050 04:05:23,820 --> 04:05:30,594 STRUCTURE AND INSTEAD OF RUNNING 5051 04:05:30,594 --> 04:05:34,331 A SIMULATION I CAN USE THIS AND 5052 04:05:34,331 --> 04:05:40,170 THE CYCLE ENDS UP BEING ZERO 5053 04:05:40,170 --> 04:05:41,571 INSTEAD DO TRANSFORMATION OF THE 5054 04:05:41,571 --> 04:05:48,812 BIND BEING SITE AND SOLVENT. 5055 04:05:48,812 --> 04:05:52,149 AND YOU CAN LOOK AT THE BEST 5056 04:05:52,149 --> 04:05:53,116 PRACTICES FOR THE LIVING JOURNAL 5057 04:05:53,116 --> 04:05:56,186 OF COMPUTATIONAL SCIENCES. 5058 04:05:56,186 --> 04:05:58,288 THE BEST PRACTICES REVIEW THAT 5059 04:05:58,288 --> 04:06:03,360 CAPTURES A SNAP SHOT OF WHAT THE 5060 04:06:03,360 --> 04:06:04,127 FIELD BELIEVES IS IMPORTANT. 5061 04:06:04,127 --> 04:06:08,765 IF YOU WANT TO LEARN MORE ABOUT 5062 04:06:08,765 --> 04:06:12,369 HOW WE DO THIS. 5063 04:06:12,369 --> 04:06:19,276 THE SAME ENVIRONMENT AND 5064 04:06:19,276 --> 04:06:22,512 SUBTRACTING THEM AND I'VE GIVEN 5065 04:06:22,512 --> 04:06:25,315 EXAMPLES OF THE CATEGORIES AND 5066 04:06:25,315 --> 04:06:27,651 CAN ALSO DRIVE ACTIVITY IF 5067 04:06:27,651 --> 04:06:29,619 YOU'RE LOOKING AT RELATED ACTIVE 5068 04:06:29,619 --> 04:06:32,756 SITES AND KINASES THAT SHARE 5069 04:06:32,756 --> 04:06:34,191 HEMOLOGY OR PREDICTING THE 5070 04:06:34,191 --> 04:06:38,795 IMPACT OF MUTATION OF TARGET. 5071 04:06:38,795 --> 04:06:40,797 WE'LL SEE PATIENT INDIVIDUALIZED 5072 04:06:40,797 --> 04:06:43,600 MUTATIONS THAT MAY CHANGE DRUG 5073 04:06:43,600 --> 04:06:45,469 RESPONSE OR SENSITIZE THE TARGET 5074 04:06:45,469 --> 04:06:46,970 TO PARTICULAR DRUGS. 5075 04:06:46,970 --> 04:06:51,408 OTHERS HAVE USED ITS FOR 5076 04:06:51,408 --> 04:06:52,175 BIOLOGICS OPTIMIZING STABILITY 5077 04:06:52,175 --> 04:06:53,643 BY INCREASING THE MELTING 5078 04:06:53,643 --> 04:06:55,912 TEMPERATURE OF A PARTICULAR TAR 5079 04:06:55,912 --> 04:06:56,079 FET. 5080 04:06:56,079 --> 04:06:59,115 THERE'S OTHER THINGS LESS TARGET 5081 04:06:59,115 --> 04:07:04,988 CENTRIC AND MAYBE ANTI-TARGET 5082 04:07:04,988 --> 04:07:10,093 BASED AND YOU AVOID CAUSING 5083 04:07:10,093 --> 04:07:10,794 CARDIO TOXICITY. 5084 04:07:10,794 --> 04:07:12,162 WE'RE USING THE CALCULATIONS TO 5085 04:07:12,162 --> 04:07:15,398 ADDRESS QUESTIONS OF ANTI-TARGET 5086 04:07:15,398 --> 04:07:17,100 BINDING AND OTHER PROPERTIES. 5087 04:07:17,100 --> 04:07:20,804 AND OTHER PHYSICAL PROPERTIES 5088 04:07:20,804 --> 04:07:22,873 LIKE SOLUBILITY AND THE SAME 5089 04:07:22,873 --> 04:07:23,707 TECHNOLOGY CAN BE USED TO 5090 04:07:23,707 --> 04:07:25,242 ADDRESS A LOT OF THESE 5091 04:07:25,242 --> 04:07:25,509 QUESTIONS. 5092 04:07:25,509 --> 04:07:30,046 THE WAY THE CALCULATIONS YOU 5093 04:07:30,046 --> 04:07:34,117 PERTURB ONE LIGAND TO ANOTHER IN 5094 04:07:34,117 --> 04:07:40,824 A COMPLEX WAY AND RUNNING 5095 04:07:40,824 --> 04:07:43,326 SIMULATIONS AND ABLE TO GET A 5096 04:07:43,326 --> 04:07:45,629 GOOD ESTIMATE OF WHAT THE FREE 5097 04:07:45,629 --> 04:07:47,564 ENERGY DIFFERENCE BINDING ARE. 5098 04:07:47,564 --> 04:07:50,100 THERE'S OTHER VARIANTS FOR FREE 5099 04:07:50,100 --> 04:07:52,102 ENERGIES MORE EXTENSIVE BUT ALL 5100 04:07:52,102 --> 04:07:55,972 USE SOME WAY OF SAMPLING THE 5101 04:07:55,972 --> 04:07:57,507 CHEMICAL STATES IN SOME WAY. 5102 04:07:57,507 --> 04:07:59,609 THERE'S DIFFERENT METHODOLOGIES 5103 04:07:59,609 --> 04:08:02,479 AND THIS IS A SNAPSHOT AND SOME 5104 04:08:02,479 --> 04:08:04,814 VIEWED AS THE BEST GOLD STANDARD 5105 04:08:04,814 --> 04:08:08,485 BUT IT'S A LITTLE BIT TRICKY TO 5106 04:08:08,485 --> 04:08:08,818 IMPLEMENT. 5107 04:08:08,818 --> 04:08:11,121 THERE'S OTHER VARIANTS WITH COOL 5108 04:08:11,121 --> 04:08:14,724 RECENT STATISTICAL MECHANIC 5109 04:08:14,724 --> 04:08:16,092 RECOVERIES THAT ALLOWS TO 5110 04:08:16,092 --> 04:08:16,793 PARALYZE THE CALCULATIONS OVER 5111 04:08:16,793 --> 04:08:16,927 MANY 5112 04:08:20,864 --> 04:08:27,304 GPUs AND FOLKS EFFECTIVELY USING 5113 04:08:27,304 --> 04:08:28,805 THEM NOW. 5114 04:08:28,805 --> 04:08:31,007 THEY ARE DON'T HAVE TO BE 5115 04:08:31,007 --> 04:08:32,542 TREMENDOUSLY ACCURATE. 5116 04:08:32,542 --> 04:08:36,813 WE'RE MAKING MODIFICATIONS AND 5117 04:08:36,813 --> 04:08:38,915 THEY'RE MODELLED IN TERMS OF 5118 04:08:38,915 --> 04:08:41,117 BINDING FREE ENERGY CHANGES. 5119 04:08:41,117 --> 04:08:43,653 IF YOU LOOK AT THE SHADED REGION 5120 04:08:43,653 --> 04:08:45,388 IT'S THE PROBABILITY OF 5121 04:08:45,388 --> 04:08:49,893 IMPROVING BY A FACTOR BY 10 AND 5122 04:08:49,893 --> 04:08:51,027 DECREASE THE COMPOUNDS TO HIT 5123 04:08:51,027 --> 04:08:53,763 THE GOAL BY A FACTOR OF 3. 5124 04:08:53,763 --> 04:08:57,267 IF YOU CAN GET TO HERE YOU ARE 5125 04:08:57,267 --> 04:08:59,135 SPEEDING THINGS REDUCING THE 5126 04:08:59,135 --> 04:09:03,473 DESIGN CYCLES BY AN ORDER OF 5127 04:09:03,473 --> 04:09:03,740 MAGNITUDE. 5128 04:09:03,740 --> 04:09:06,276 THERE'S EFFORTS THAT GO THROUGH 5129 04:09:06,276 --> 04:09:10,680 AND USING TOOL TO COMPUTE 5130 04:09:10,680 --> 04:09:12,048 RETROSPECTIVE BENCH MARKS AND 5131 04:09:12,048 --> 04:09:16,519 CURATED AND CAN ASSESS THE 5132 04:09:16,519 --> 04:09:19,923 ACCURACY ON TARGET CLASSES AN 5133 04:09:19,923 --> 04:09:20,790 SCAFF 5134 04:09:20,790 --> 04:09:21,057 SCAFFOLDS. 5135 04:09:21,057 --> 04:09:22,292 SOMETIMES IT WORKS WELL AND 5136 04:09:22,292 --> 04:09:24,094 SOMETIMES NOT AS WELL. 5137 04:09:24,094 --> 04:09:28,798 OVER ALL THE ACCURACY IS PRETTY 5138 04:09:28,798 --> 04:09:30,934 GOOD. 5139 04:09:30,934 --> 04:09:31,968 IT'S EXCEPTIONALLY USEFUL IN 5140 04:09:31,968 --> 04:09:34,871 PRIORITIZING WHICH COMPOUNDS YOU 5141 04:09:34,871 --> 04:09:37,507 MAKE AND GETTING RID OF BAND 5142 04:09:37,507 --> 04:09:38,608 DESIGNS AND ADDRESS THE 5143 04:09:38,608 --> 04:09:40,810 QUESTIONS WE'RE ANSWERING OR 5144 04:09:40,810 --> 04:09:45,949 ASKING ABOUT BINDING AFFINITY IN 5145 04:09:45,949 --> 04:09:46,950 THE PUBLICATION. 5146 04:09:46,950 --> 04:09:49,719 WE'VE USED THIS AS PART OF THE 5147 04:09:49,719 --> 04:09:51,454 COVID MOON SHOT SUCCESS. 5148 04:09:51,454 --> 04:09:53,657 THERE'S OPPORTUNITIES TO MAKE 5149 04:09:53,657 --> 04:09:59,129 IMPROVEMENTS IN VARIOUS SCAF 5150 04:09:59,129 --> 04:10:02,966 SCAFFOLDS AND THIS IS REALLY 5151 04:10:02,966 --> 04:10:05,802 WORKING CLOSELY WITH CHEMISTS 5152 04:10:05,802 --> 04:10:07,303 DRIVING THE IDEATION AND 5153 04:10:07,303 --> 04:10:08,438 DECISION MAKING IN EACH STAGE 5154 04:10:08,438 --> 04:10:10,440 AND IT TAKES A TEAM TO MAKE ALL 5155 04:10:10,440 --> 04:10:11,474 THIS HAPPEN. 5156 04:10:11,474 --> 04:10:13,810 YOU CAN FIND ALL THE DATA AND 5157 04:10:13,810 --> 04:10:16,746 READ ABOUT THE STORY ALL OPEN 5158 04:10:16,746 --> 04:10:16,980 SCIENCE. 5159 04:10:16,980 --> 04:10:18,148 SOMETIMES HOWEVER, THINGS DID 5160 04:10:18,148 --> 04:10:19,015 NOT GO SO WELL. 5161 04:10:19,015 --> 04:10:22,552 WE FREQUENTLY FOUND WE HAD 5162 04:10:22,552 --> 04:10:25,422 MISMODELLED THE ACTIVE SITE AND 5163 04:10:25,422 --> 04:10:27,457 FOUND SOME COMPOUNDS AGREE WELL 5164 04:10:27,457 --> 04:10:29,025 AND OTHERS HAVE NO CORRELATION 5165 04:10:29,025 --> 04:10:30,660 BETWEEN EXPERIMENT CALCULATION. 5166 04:10:30,660 --> 04:10:35,432 THIS IS A MORE GENERAL 5167 04:10:35,432 --> 04:10:35,699 PHENOMENA. 5168 04:10:35,699 --> 04:10:36,833 THERE'S A PAPER I ENCOURAGE YOU 5169 04:10:36,833 --> 04:10:39,269 TO CHECK OUT ABOUT THE KINDS OF 5170 04:10:39,269 --> 04:10:41,337 SUCCESSES AND FAILURES FOUND IN 5171 04:10:41,337 --> 04:10:44,374 TRYING TO APPLY FREE ENERGY 5172 04:10:44,374 --> 04:10:46,509 CALCULATIONS TO PROGRAMS 5173 04:10:46,509 --> 04:10:47,744 INTERNALLY AT MERCK AND 5174 04:10:47,744 --> 04:10:49,979 CATEGORIZED THE AREAS AND 5175 04:10:49,979 --> 04:10:52,816 CLASSES OF FAILURES OBSERVED AND 5176 04:10:52,816 --> 04:10:54,984 WAS ABLE TO PIN DOWN. 5177 04:10:54,984 --> 04:10:55,985 IN TOTAL HALF THE PROGRAMS 5178 04:10:55,985 --> 04:10:57,721 WORKED WELL. 5179 04:10:57,721 --> 04:11:01,257 HALF WERE FAILURES FOR DIFFERENT 5180 04:11:01,257 --> 04:11:01,558 REASONS. 5181 04:11:01,558 --> 04:11:02,892 ONE OTHER IMPORTANT REASON IS 5182 04:11:02,892 --> 04:11:06,129 THE CURRENT LEVEL OF ACCURACY IS 5183 04:11:06,129 --> 04:11:07,197 AN OPPORTUNITY TO IMPROVE. 5184 04:11:07,197 --> 04:11:09,466 WITH CAN'T RUN HUNDREDS OF 5185 04:11:09,466 --> 04:11:11,134 PROGRAMS OVER AND OVER AND SEE 5186 04:11:11,134 --> 04:11:13,169 WHAT THE IMPACT OF MAKING A TOOL 5187 04:11:13,169 --> 04:11:16,806 IS ON THE DISTRIBUTION FUNCTION 5188 04:11:16,806 --> 04:11:22,312 BUT CAN DO THIS SYNTHETICALLY 5189 04:11:22,312 --> 04:11:25,248 AND MODEL PERHAPS IF WE USE 5190 04:11:25,248 --> 04:11:27,417 TOOLS OF DIFFERENT ACCURACY AND 5191 04:11:27,417 --> 04:11:30,019 HOW IT SHIFTS COSTS IN A MODEL 5192 04:11:30,019 --> 04:11:33,590 PHASE OF THE REALISTIC PROGRAM. 5193 04:11:33,590 --> 04:11:36,760 YOU SEE THE AMOUNT OF MONEY IT'S 5194 04:11:36,760 --> 04:11:38,094 SAVED GETTING FASTER AND CHEAPER 5195 04:11:38,094 --> 04:11:40,196 CAN BE ENORMOUS IN REDUCING THE 5196 04:11:40,196 --> 04:11:40,764 MODELLING ERROR. 5197 04:11:40,764 --> 04:11:42,732 THERE'S INTEREST IN MAKING THEM 5198 04:11:42,732 --> 04:11:45,835 MORE ACCURATE BY IMPROVING ON 5199 04:11:45,835 --> 04:11:48,338 THEIR PERFORMANCE. 5200 04:11:48,338 --> 04:11:50,940 THERE'S THREE CRITICAL AREAS. 5201 04:11:50,940 --> 04:11:55,411 ONE IS THE ACCURACY OR 5202 04:11:55,411 --> 04:11:56,813 SIMULATION MELD AND THE ITEMS IN 5203 04:11:56,813 --> 04:11:58,081 THE SYSTEM. 5204 04:11:58,081 --> 04:12:01,050 WE'VE USED THE FORCE FIELD FORMS 5205 04:12:01,050 --> 04:12:02,986 ABOUT 4R5 YEARS OLD NOW. 5206 04:12:02,986 --> 04:12:04,788 THERE'S AMPLE OPPORTUNITY TO 5207 04:12:04,788 --> 04:12:07,557 MORE TO MORE FASTER HARDWARE AND 5208 04:12:07,557 --> 04:12:12,762 LOOKING AT THE WAY TO IMPROVE 5209 04:12:12,762 --> 04:12:15,832 AND NEGLECT EFFECTS THAT CAN BE 5210 04:12:15,832 --> 04:12:19,068 IMPORTANT IN PARTICULAR 5211 04:12:19,068 --> 04:12:20,804 INTERACTION AND INTEREST IN 5212 04:12:20,804 --> 04:12:21,437 MODELLING CONFIRMATIONS IN 5213 04:12:21,437 --> 04:12:24,741 PROTEIN RELEVANT AND PROTEIN 5214 04:12:24,741 --> 04:12:27,577 LIGANDS THAT MAY BE RELEVANT AND 5215 04:12:27,577 --> 04:12:29,312 OFTEN THERE'S MILLISECOND SCALE 5216 04:12:29,312 --> 04:12:32,749 AND THE MOTIONS ARE SHORTER AT 5217 04:12:32,749 --> 04:12:33,850 NANOSECONDS. 5218 04:12:33,850 --> 04:12:39,455 ALL RIPE FOR A.I., M.L. TO 5219 04:12:39,455 --> 04:12:40,557 ACCELERATE PROGRESS TO MAKE IT 5220 04:12:40,557 --> 04:12:40,790 BETTER. 5221 04:12:40,790 --> 04:12:44,828 FIRST YOU START WITH A GOOD OPEN 5222 04:12:44,828 --> 04:12:46,896 FOUNDATION AS A BASELINE AND 5223 04:12:46,896 --> 04:12:47,964 INTEGRATIVE TECHNOLOGIES AND 5224 04:12:47,964 --> 04:12:49,199 ASSESSING THEM IN DIFFERENT 5225 04:12:49,199 --> 04:12:49,666 WAYS. 5226 04:12:49,666 --> 04:12:52,535 THE SOFTWARE FOUNDATION IS A 5227 04:12:52,535 --> 04:12:54,404 NON-PROFIT BUILDING THE WHOLE 5228 04:12:54,404 --> 04:12:56,806 ECO SYSTEM OF OPEN SOURCE TOOLS 5229 04:12:56,806 --> 04:13:00,476 FOR ACCELERATING DRUG DISCOVERY. 5230 04:13:00,476 --> 04:13:03,847 IT'S FROM THE OPEN FORCE FIELDS 5231 04:13:03,847 --> 04:13:06,015 ACCORDING TO BEST PRACTICES AND 5232 04:13:06,015 --> 04:13:11,087 OPEN FREE ENERGY AND PRODUCING 5233 04:13:11,087 --> 04:13:13,923 MODELS AN AIMING TO PREDICT 5234 04:13:13,923 --> 04:13:16,626 PROPERTIES USING NEW DATA FUNDED 5235 04:13:16,626 --> 04:13:18,294 BY AN ARPA-H GRANT. 5236 04:13:18,294 --> 04:13:20,530 THERE'S OPEN SOURCE TOOLS AND 5237 04:13:20,530 --> 04:13:22,465 DATA SETS AS WELL AS OPEN 5238 04:13:22,465 --> 04:13:24,667 ITERATIONS OF FORCE FOLDS THAT 5239 04:13:24,667 --> 04:13:27,971 CONFORM TO BEST PRACTICES. 5240 04:13:27,971 --> 04:13:33,877 YOU CAN FIND MORE BUT I WANT TO 5241 04:13:33,877 --> 04:13:35,144 HIGHLIGHT THE OPEN ENERGY 5242 04:13:35,144 --> 04:13:38,047 CONSORTIUM AND EXPERIMENT HOW TO 5243 04:13:38,047 --> 04:13:41,818 DO BETTER USING OTHER M.L. 5244 04:13:41,818 --> 04:13:42,051 METHODS. 5245 04:13:42,051 --> 04:13:44,954 THERE'S A STAFF OF PROFESSIONAL 5246 04:13:44,954 --> 04:13:46,923 DEVELOPERS CONTRIBUTING AND 5247 04:13:46,923 --> 04:13:49,926 COMMITTEE OF AM DEMOCRATIC 5248 04:13:49,926 --> 04:13:50,860 MEMBERS AND GOVERNANCE AND 5249 04:13:50,860 --> 04:13:55,331 GUIDANCE AND FUNDING TO MAKE THE 5250 04:13:55,331 --> 04:13:56,232 EFFORT SUSTAINABLE. 5251 04:13:56,232 --> 04:13:59,135 THEY'VE BEEN WORKING ON BEST 5252 04:13:59,135 --> 04:14:01,504 PRACTICES CAPTURED FROM A BROAD 5253 04:14:01,504 --> 04:14:02,805 INDUSTRY CROSS SECTION IN THE 5254 04:14:02,805 --> 04:14:04,807 PAPERS AND WHAT'S COOL IS YOU 5255 04:14:04,807 --> 04:14:10,380 CAN TRY THIS IN A GOOGLE 5256 04:14:10,380 --> 04:14:12,282 NOTEBOOK AND PLAY WITH THE 5257 04:14:12,282 --> 04:14:12,815 METHODS TODAY. 5258 04:14:12,815 --> 04:14:15,585 THERE'S TOOLS AS PART OF THIS 5259 04:14:15,585 --> 04:14:15,952 INFRASTRUCTURE. 5260 04:14:15,952 --> 04:14:17,153 THERE'S TOOLS FOR PLANNING 5261 04:14:17,153 --> 04:14:18,288 CAMPAIGNS OF TRANSFORMATION 5262 04:14:18,288 --> 04:14:23,126 NETWORKS WHERE YOU MAY HAVE 5263 04:14:23,126 --> 04:14:29,532 RELATED DESIGNS AND ITERATING 5264 04:14:29,532 --> 04:14:31,935 TRANSFORMATIONS AND THERE'S AN 5265 04:14:31,935 --> 04:14:34,404 ALSO OPEN SOURCE TOOLS THAT LOU 5266 04:14:34,404 --> 04:14:36,539 TO YOU EXECUTE THE CALCULATIONS 5267 04:14:36,539 --> 04:14:38,174 RAPIDLY AND CURRENTLY TAKE 12 5268 04:14:38,174 --> 04:14:40,677 HOURS FOR TRANSFORMATION SO 5269 04:14:40,677 --> 04:14:43,780 THEY'RE QUITE EXTENSIVE COMPARED 5270 04:14:43,780 --> 04:14:46,416 TO M.L. APPROACHES. 5271 04:14:46,416 --> 04:14:48,818 THERE'S DIFFERENT ALGORITHMS 5272 04:14:48,818 --> 04:14:54,557 THAT ALLOW US TO BE RAPIDLY 5273 04:14:54,557 --> 04:14:55,892 EXTENDED AND YOU CAN IMPROVE AND 5274 04:14:55,892 --> 04:14:58,861 SWAP OUT A MODULE AND DO BETTER 5275 04:14:58,861 --> 04:15:00,797 AS WELL AS USING BEST PRACTICES 5276 04:15:00,797 --> 04:15:02,799 AND ANALYZING THE CALCULATION 5277 04:15:02,799 --> 04:15:04,233 AND REPRESENTING HOW YOU ASSESS 5278 04:15:04,233 --> 04:15:05,635 THE CALCULATIONS IN A BEST 5279 04:15:05,635 --> 04:15:07,136 PRACTICES MANNER TO MAKE SURE 5280 04:15:07,136 --> 04:15:11,074 YOU'RE NOT BEING DECEIVED BY 5281 04:15:11,074 --> 04:15:11,374 STATISTICS. 5282 04:15:11,374 --> 04:15:12,942 THERE'S TOOLS AVAILABLE FOR 5283 04:15:12,942 --> 04:15:14,610 ORCHESTRATING LARGE CAMPAIGNS OF 5284 04:15:14,610 --> 04:15:16,813 THESE ON DISTRIBUTED OR CLOUD 5285 04:15:16,813 --> 04:15:17,547 COMPUTING RESOURCES. 5286 04:15:17,547 --> 04:15:19,682 WE MANAGED TO USE THESE BETWEEN 5287 04:15:19,682 --> 04:15:22,785 THE ANTIVIRAL DISCOVERY CENTER 5288 04:15:22,785 --> 04:15:26,155 SAP AND TWO PROJECTS IN THE LAST 5289 04:15:26,155 --> 04:15:27,457 YEAR ALONE. 5290 04:15:27,457 --> 04:15:29,692 THESE ARE GETTING TO BE ROBUST 5291 04:15:29,692 --> 04:15:30,860 AND SCALABLE. 5292 04:15:30,860 --> 04:15:32,795 THE AMAZING THING IS THEY'RE 5293 04:15:32,795 --> 04:15:34,364 ABLE TO DO CONTINUOUS 5294 04:15:34,364 --> 04:15:35,698 BENCHMARKING USING RESOURCES TO 5295 04:15:35,698 --> 04:15:36,799 MAKE SURE THEY'RE ALWAYS 5296 04:15:36,799 --> 04:15:38,234 IMPROVING AS THEY ADD NEW 5297 04:15:38,234 --> 04:15:39,669 FEATURES OR CHANGE THE ECO 5298 04:15:39,669 --> 04:15:41,938 SYSTEM AND MAKE SURE THEY 5299 04:15:41,938 --> 04:15:43,473 HAVEN'T BROKEN SOMETHING AND 5300 04:15:43,473 --> 04:15:44,340 THERE'S EFFORTS DELIVERING THE 5301 04:15:44,340 --> 04:15:46,175 BEST PRACTICES TO YOU AS PART OF 5302 04:15:46,175 --> 04:15:50,146 WHAT'S GOING ON IN THE TOOLS. 5303 04:15:50,146 --> 04:15:51,080 YOU CAN BUILD PROFESSIONAL 5304 04:15:51,080 --> 04:15:54,617 QUALITY DRUG DISCOVERY ECO 5305 04:15:54,617 --> 04:15:55,251 SYSTEMS. 5306 04:15:55,251 --> 04:16:00,356 JOSH HORTON AND HUGO AS PART OF 5307 04:16:00,356 --> 04:16:01,758 THE ANTIVIRAL DRUG DISCOVERY 5308 04:16:01,758 --> 04:16:03,259 CENTER WE LEAD THE CENTERS 5309 04:16:03,259 --> 04:16:04,794 WORKING ON PANDEMIC 5310 04:16:04,794 --> 04:16:05,128 PREPAREDNESS. 5311 04:16:05,128 --> 04:16:08,798 THIS IS THE ENTIRE WORK FLOW YOU 5312 04:16:08,798 --> 04:16:11,868 CAN DOWNLOAD AND ADAPT TO YOUR 5313 04:16:11,868 --> 04:16:14,904 OWN DRUG DISCOVERY NEEDS BY 5314 04:16:14,904 --> 04:16:16,572 GOING TO GITHUB. 5315 04:16:16,572 --> 04:16:18,241 EVERYTHING IS OPEN SOURCE. 5316 04:16:18,241 --> 04:16:20,043 INDUSTRY IS REPORTING SUCCESSFUL 5317 04:16:20,043 --> 04:16:23,346 USE OF THESE OPEN SOURCE 5318 04:16:23,346 --> 04:16:25,715 PIPELINES WITH REASONABLE 5319 04:16:25,715 --> 04:16:26,215 ACCURACY. 5320 04:16:26,215 --> 04:16:26,716 WE'VE. 5321 04:16:26,716 --> 04:16:29,152 USING THIS IN OUR NEW ANTIVIRAL 5322 04:16:29,152 --> 04:16:30,486 DISCOVERY PROGRAMS BUILDING 5323 04:16:30,486 --> 04:16:38,628 SECOND AND THIRD GENERATION 5324 04:16:38,628 --> 04:16:40,063 INTERERS FOR SARS COV2 AND 5325 04:16:40,063 --> 04:16:42,398 BEYOND AND WE LOOK AT 5326 04:16:42,398 --> 04:16:44,367 SUSTAINABLE FUNDING. 5327 04:16:44,367 --> 04:16:46,536 IT'S A GREAT FEATURE OF THE ECO 5328 04:16:46,536 --> 04:16:48,738 SYSTEM AND THERE'S A HUGE NUMBER 5329 04:16:48,738 --> 04:16:50,273 OF FOLKS THAT CONTRIBUTED 5330 04:16:50,273 --> 04:16:51,641 SCIENTIFICALLY AND OTHERWISE. 5331 04:16:51,641 --> 04:16:53,109 WITH THAT I'LL REMIND YOU MY 5332 04:16:53,109 --> 04:16:55,878 SLIDES ARE AVAILABLE HERE IF YOU 5333 04:16:55,878 --> 04:16:58,114 WANT TO LOOK AT ANY AND I'M 5334 04:16:58,114 --> 04:16:59,982 HAPPY TO TAKE ANY QUESTIONS. 5335 04:16:59,982 --> 04:17:04,620 I HOPE I KEPT US ON TIME. 5336 04:17:04,620 --> 04:17:06,389 >> YOU KEPT AHEAD OF TIME. 5337 04:17:06,389 --> 04:17:08,791 >> EXCELLENT. 5338 04:17:08,791 --> 04:17:17,834 >> THANK YOU VERY MUCH. 5339 04:17:17,834 --> 04:17:21,471 YOU SPOKE FAST. 5340 04:17:21,471 --> 04:17:31,114 A FEW QUESTIONS FROM THE Q&A. 5341 04:17:31,114 --> 04:17:35,017 YOU HAVE ALLUDED TO USING 5342 04:17:35,017 --> 04:17:40,790 MACHINE LEARNING AND MAYBE GIVEN 5343 04:17:40,790 --> 04:17:47,997 WE HAVE TIME HOW YOU SEE THE 5344 04:17:47,997 --> 04:17:52,969 METHOD GOING FORWARD. 5345 04:17:52,969 --> 04:17:54,537 THERE'S SEVERAL LEVELS. 5346 04:17:54,537 --> 04:17:56,272 THE FIRST IS THESE CAMPAIGNS ARE 5347 04:17:56,272 --> 04:17:57,273 MORE SUCCESSFUL THE MORE 5348 04:17:57,273 --> 04:17:58,374 COMPOUNDS YOU SCORE. 5349 04:17:58,374 --> 04:18:00,776 ONE OF THE SIMPLE AREAS IS TO 5350 04:18:00,776 --> 04:18:04,814 AVOID THE COST OF 12GPU HOURS 5351 04:18:04,814 --> 04:18:09,385 PER TRANSFORMATION CAN WE SCORE 5352 04:18:09,385 --> 04:18:14,924 LARGER SETS FOR A NEW TARGET 5353 04:18:14,924 --> 04:18:18,127 WITHOUT EVERY LIGAND 5354 04:18:18,127 --> 04:18:20,429 TRANSFORMATION AND THERE'S 3-D 5355 04:18:20,429 --> 04:18:22,064 AND 2-D METHODS. 5356 04:18:22,064 --> 04:18:25,968 THIS WILL SAVE HUGE AMOUNTS OF 5357 04:18:25,968 --> 04:18:26,903 CLOUD COMPUTING TIME. 5358 04:18:26,903 --> 04:18:27,436 THAT'S ONE OPPORTUNITY. 5359 04:18:27,436 --> 04:18:32,441 ANOTHER IS TO LOOK AT EACH AREA 5360 04:18:32,441 --> 04:18:33,709 WHERE THERE'S NEED TO IMPROVE. 5361 04:18:33,709 --> 04:18:38,147 FOLKS ARE, WOULD GO TO TURN THIS 5362 04:18:38,147 --> 04:18:38,814 PROBLEM INTO A MACHINE LEARNING 5363 04:18:38,814 --> 04:18:40,216 PROBLEM TO LEARN FROM DATA AND 5364 04:18:40,216 --> 04:18:42,518 BE TRAINED ON LARGE DATA SETS. 5365 04:18:42,518 --> 04:18:44,820 THERE'S A MOVEMENT TOWARDS 5366 04:18:44,820 --> 04:18:46,956 BUILDING FOUNDATION MODELS USING 5367 04:18:46,956 --> 04:18:52,562 MORE SFOPHISTICATED FORMS LIKE 5368 04:18:52,562 --> 04:18:56,499 NEURAL NETWORKS AND HOW TO DO A 5369 04:18:56,499 --> 04:18:58,935 BETTER INTEGRATION OF 5370 04:18:58,935 --> 04:19:03,072 STATISTICAL MECHANICS WITH 5371 04:19:03,072 --> 04:19:05,875 MACHINE LEARNING APPROACHES AND 5372 04:19:05,875 --> 04:19:07,610 THERE'S INTEREST IN ALL THE 5373 04:19:07,610 --> 04:19:09,812 VARIANTS THAT WILL PREDICT 5374 04:19:09,812 --> 04:19:12,815 MULTIPLE PROTEIN CONFIRMATIONS 5375 04:19:12,815 --> 04:19:18,387 AND IDEALLY LIGANDS INTO WAYS TO 5376 04:19:18,387 --> 04:19:22,158 SAMPLE FROM THE RIGHT DENSITY IS 5377 04:19:22,158 --> 04:19:23,226 A HUGE OPEN OPPORTUNITY TO 5378 04:19:23,226 --> 04:19:24,794 ACCELERATE THE ACCURACY AND 5379 04:19:24,794 --> 04:19:26,262 RELIABILITY OF THE CAL CLAYINGS. 5380 04:19:26,262 --> 04:19:31,500 THERE'S AMPLE AREA IN MANY 5381 04:19:31,500 --> 04:19:33,202 PLACES FOR MACHINE LEARNING TO 5382 04:19:33,202 --> 04:19:34,303 SOLVE THE PROBLEMS. 5383 04:19:34,303 --> 04:19:39,709 >> GREAT USE OF MACHINE LEARNING 5384 04:19:39,709 --> 04:19:41,143 IN THE AREA THAT HAVE 5385 04:19:41,143 --> 04:19:45,114 HISTORICALLY BE ANYONE UNDER 5386 04:19:45,114 --> 04:19:45,548 RESOURCED. 5387 04:19:45,548 --> 04:19:46,082 >> IT SEEMS LIKE A MISSED 5388 04:19:46,082 --> 04:19:56,292 OPPORTUNITY. 5389 04:20:00,329 --> 04:20:05,067 THERE'S SYSTEMS WHERE IT'S NOT 5390 04:20:05,067 --> 04:20:05,868 AVAILABLE. 5391 04:20:05,868 --> 04:20:07,403 >> THERE'S SOME WORK OUT THERE. 5392 04:20:07,403 --> 04:20:11,140 YOU'LL SEE IN RECENT PAPERS 5393 04:20:11,140 --> 04:20:19,348 TRYING TO USE ALPHA FOLD AND 5394 04:20:19,348 --> 04:20:22,618 THERE'S PROBLEMS MAKING IT FIT 5395 04:20:22,618 --> 04:20:25,388 AND YOU SOMETIMES HAVE TO USE 5396 04:20:25,388 --> 04:20:27,323 DOCKING AND THERE'S METHODS THAT 5397 04:20:27,323 --> 04:20:31,027 PREDICT THE STRUCTURE. 5398 04:20:31,027 --> 04:20:32,795 THIS IS WHERE MOST ENTHUSIASM IS 5399 04:20:32,795 --> 04:20:34,130 THEN YOU'RE OFF TO THE RACES 5400 04:20:34,130 --> 04:20:35,464 WITH THE FREE ENERGY 5401 04:20:35,464 --> 04:20:35,798 CALCULATIONS. 5402 04:20:35,798 --> 04:20:44,807 THE ANSWER IS WE DON'T KNOW YET. 5403 04:20:44,807 --> 04:20:47,576 NOT EXHAUSTIVELY BUT THERE'S 5404 04:20:47,576 --> 04:20:50,313 DOCKING TO ALPHA FOLD PREDICTIVE 5405 04:20:50,313 --> 04:20:56,585 STRUCTURES WHERE THERE'S BEEN 5406 04:20:56,585 --> 04:20:58,821 SIGNIFICANT SUCCESS. 5407 04:20:58,821 --> 04:21:04,260 >> HERE'S A LOADED QUESTION. 5408 04:21:04,260 --> 04:21:06,162 THE SHORT VERSION IS DOES IT 5409 04:21:06,162 --> 04:21:08,898 SEEM NO COMMERCIAL TOOLS ARE 5410 04:21:08,898 --> 04:21:11,667 NEEDED FOR DRUG DISCOVERY. 5411 04:21:11,667 --> 04:21:17,373 MY EXTENSION TO THE QUESTION IS 5412 04:21:17,373 --> 04:21:19,975 HOW DO WE PLAY WITH INDUSTRY AND 5413 04:21:19,975 --> 04:21:25,147 OPEN WORK, WHO FUNDS IT AND HOW 5414 04:21:25,147 --> 04:21:29,051 DO WE MAKE THE TOOLS VALIDATIVE, 5415 04:21:29,051 --> 04:21:34,156 USEFUL AND FUNDED? 5416 04:21:34,156 --> 04:21:40,563 >> GREAT QUESTION. 5417 04:21:40,563 --> 04:21:44,834 THE HOOK IS INDUSTRY WILL HAVE 5418 04:21:44,834 --> 04:21:46,435 FIRST THERE'S A LOT OF UTILITY 5419 04:21:46,435 --> 04:21:48,804 FOR INDUSTRY PRODUCED SOFTWARE 5420 04:21:48,804 --> 04:21:50,906 IN HOUSE AND THROUGH LARGE 5421 04:21:50,906 --> 04:21:52,041 COMMERCIAL SOFTWARE VENDORS BUT 5422 04:21:52,041 --> 04:21:53,509 NOT EVERYONE NEEDS TO INVENT 5423 04:21:53,509 --> 04:21:54,477 EVERYTHING FROM SCRATCH. 5424 04:21:54,477 --> 04:21:58,214 THERE'S NO NEED TO HAVE AN 5425 04:21:58,214 --> 04:22:01,951 ENTIRE ECO SYSTEM FROM SCRATCH. 5426 04:22:01,951 --> 04:22:05,121 THE HOOK IS TO HAVE A GOOD SOLID 5427 04:22:05,121 --> 04:22:08,791 FOUNDATION WHICH WE THINK OF AS 5428 04:22:08,791 --> 04:22:13,596 THE PARTS OF THE OPEN SOURCE ECO 5429 04:22:13,596 --> 04:22:14,830 SYSTEM AND START WITH BEST 5430 04:22:14,830 --> 04:22:16,732 PRACTICE AND INNOVATE WHERE YOU 5431 04:22:16,732 --> 04:22:18,834 WANT TO HAVE THE PEST IDEAS OR 5432 04:22:18,834 --> 04:22:20,436 HAVE THE HIGHEST OR BEST 5433 04:22:20,436 --> 04:22:23,572 CONTRIBUTIONS OR WANT TO FOCUS 5434 04:22:23,572 --> 04:22:24,807 YOUR ENERGY. 5435 04:22:24,807 --> 04:22:26,675 AND EVERYTHING IS MODULAR. 5436 04:22:26,675 --> 04:22:28,677 YOU CAN TAKE THE FREE ENERGY 5437 04:22:28,677 --> 04:22:31,614 TOOL AND REPLACE THE PROTOCOL 5438 04:22:31,614 --> 04:22:34,049 OBJECT WITH YOUR FAVORITE FREE 5439 04:22:34,049 --> 04:22:36,552 ENERGY METHOD AND DROP INTO THE 5440 04:22:36,552 --> 04:22:38,821 ECHO SYSTEM WITH OTHER BEST 5441 04:22:38,821 --> 04:22:39,355 PRACTICES METHODS. 5442 04:22:39,355 --> 04:22:41,090 THERE'S OPPORTUNITY FOR SYNERGY 5443 04:22:41,090 --> 04:22:43,259 AND FOR SAVING INDUSTRY A LOT OF 5444 04:22:43,259 --> 04:22:45,161 TIME IN AREAS WHERE IT DOESN'T 5445 04:22:45,161 --> 04:22:46,695 NEED TO ENGINEER EVERYTHING. 5446 04:22:46,695 --> 04:22:47,797 THE CURRENT MODEL WE FOCUSSED ON 5447 04:22:47,797 --> 04:22:50,466 RIGHT NOW IS WHERE WE HAVE 5448 04:22:50,466 --> 04:22:53,869 INDUSTRY FUNDED CONSORTIA WITH 5449 04:22:53,869 --> 04:22:55,771 ACADEMIC-LED FEDERAL FUNDING OR 5450 04:22:55,771 --> 04:22:58,541 NONPROFIT FUNDING SUPPORTING 5451 04:22:58,541 --> 04:22:59,809 THESE PROJECTS JOINTLY. 5452 04:22:59,809 --> 04:23:04,113 THAT MODEL HAS BEEN SUCCESSFUL 5453 04:23:04,113 --> 04:23:05,581 FOR THE OPEN FORCE FIELD 5454 04:23:05,581 --> 04:23:07,516 INITIATIVE WHICH HAS GROWN UNDER 5455 04:23:07,516 --> 04:23:11,120 THAT FUNDING AND SCIENTIFIC 5456 04:23:11,120 --> 04:23:14,457 INPUT AND GOVERNANCE MODEL 5457 04:23:14,457 --> 04:23:18,527 HAVING ACADEMICS AND INDUSTRY 5458 04:23:18,527 --> 04:23:25,968 ARE MAKING TOOLS PEOPLE CAN USE. 5459 04:23:25,968 --> 04:23:28,070 >> I THINK YOU ANSWERED. 5460 04:23:28,070 --> 04:23:30,272 >> EVERYONE SHOULD FUND OPEN 5461 04:23:30,272 --> 04:23:36,278 SCIENCE PROJECTS, ALEX. 5462 04:23:36,278 --> 04:23:40,783 >> GIVING BACK USEFUL TOOLS. 5463 04:23:40,783 --> 04:23:44,620 THAT'S KIND OF WHAT YOUR ANSWER 5464 04:23:44,620 --> 04:23:46,789 IS. 5465 04:23:46,789 --> 04:23:48,424 >> I AGREE. 5466 04:23:48,424 --> 04:23:51,293 WHAT LESSONS WERE LEARNED FROM 5467 04:23:51,293 --> 04:23:53,128 THE MOON SHOT PROJECT? 5468 04:23:53,128 --> 04:23:54,497 >> WE'VE DONE THINGS DIFFERENT 5469 04:23:54,497 --> 04:23:57,433 AS PART OF THE SAP I LEAD. 5470 04:23:57,433 --> 04:23:58,534 GO TO SAP DISCOVERY AND LEARN 5471 04:23:58,534 --> 04:24:04,273 ABOUT THE CHOICES WE MADE THAT 5472 04:24:04,273 --> 04:24:04,807 ARE DIFFERENT. 5473 04:24:04,807 --> 04:24:09,678 ONE IS WE GOT INVOLVED IN 5474 04:24:09,678 --> 04:24:10,346 CROWDFUNDING BOUGHT IT'S A LOT 5475 04:24:10,346 --> 04:24:11,981 OF EFFORT AND YOU KNOW YOU HAVE 5476 04:24:11,981 --> 04:24:14,116 JOBS TO GO BACK TO. 5477 04:24:14,116 --> 04:24:17,353 WE TRIED TO AUTOMATE AS MUCH AS 5478 04:24:17,353 --> 04:24:21,223 WE CAN WITH THESE A.I./M.L. 5479 04:24:21,223 --> 04:24:23,025 METHODS AND CHECK OUT THE PAPERS 5480 04:24:23,025 --> 04:24:24,793 FROM ALPHA LEAD. 5481 04:24:24,793 --> 04:24:26,061 THE OTHER PART IS WE WENT 5482 04:24:26,061 --> 04:24:31,834 WITHOUT A PATENT AND THERE'S 5483 04:24:31,834 --> 04:24:35,938 OPPORTUNITIES FOR HAVING ENOUGH 5484 04:24:35,938 --> 04:24:36,972 GOVERNMENTAL FUNDING FOR THE 5485 04:24:36,972 --> 04:24:39,308 MODEL TO BE SUCCESSFUL BUT AS 5486 04:24:39,308 --> 04:24:40,943 WE'VE SEEN RECENTLY, GOVERNMENT 5487 04:24:40,943 --> 04:24:43,946 FUNDING FOR ANY PANDEMIC 5488 04:24:43,946 --> 04:24:46,181 PREPAREDNESS GLOBALLY HAS DRIED 5489 04:24:46,181 --> 04:24:47,116 UP. 5490 04:24:47,116 --> 04:24:53,756 THE NIH ENDED OUR ANTIVIRAL DRUG 5491 04:24:53,756 --> 04:24:54,657 DISCOVERY PROGRAMS TWO YEARS 5492 04:24:54,657 --> 04:24:56,759 EARLIER THAN EXPECTED SO IT'S 5493 04:24:56,759 --> 04:24:58,294 DIFFICULT TO PURSUE. 5494 04:24:58,294 --> 04:25:01,196 WE CAN TRY TO KEEP AFFORDABLE 5495 04:25:01,196 --> 04:25:03,732 ACCESS AND HAVE A NEW IP MODEL 5496 04:25:03,732 --> 04:25:05,734 TO MAKE SURE THE PATENT WILL 5497 04:25:05,734 --> 04:25:09,371 HAVE A SINGLE MOLECULE AND 5498 04:25:09,371 --> 04:25:12,308 LICENSED FOR NO COST AS LONG AS 5499 04:25:12,308 --> 04:25:16,779 MANUFACTURERS AGREE TO WAVE 5500 04:25:16,779 --> 04:25:23,152 THEIR EXCLUSIVITY AND THERE'S 5501 04:25:23,152 --> 04:25:32,795 MINOR TWEAKS WE WROTE ABOUT. 5502 04:25:33,529 --> 04:25:42,104 >> I I'D LIKE TO END THE 5503 04:25:42,104 --> 04:25:44,807 DISCUSSION. 5504 04:25:44,807 --> 04:25:46,875 THANK YOU FOR YOUR EXCELLENT 5505 04:25:46,875 --> 04:25:51,480 TALK AND WE FINISHED ON TIME. 5506 04:25:51,480 --> 04:25:53,816 SARINE IF YOU HAVE CONCLUDING 5507 04:25:53,816 --> 04:25:58,721 REMARKS, PLEASE PROCEED. 5508 04:25:58,721 --> 04:26:01,924 >> THANK YOU SO MUCH. 5509 04:26:01,924 --> 04:26:03,692 >> WE APPRECIATE YOU ALL. 5510 04:26:03,692 --> 04:26:06,629 I'M GOING TO BRING BACK THE 5511 04:26:06,629 --> 04:26:17,172 AGENDA SO EVERYONE KNOWS WHERE 5512 04:26:20,609 --> 04:26:20,809 WE'RE AT. 5513 04:26:20,809 --> 04:26:24,780 WE HAVE A SHORT BREAK ABOUT 10 5514 04:26:24,780 --> 04:26:28,784 MINUTES AND WE'LL BE BACK AT 5515 04:26:28,784 --> 04:26:32,788 4:15 TO START SESSION 3 ON 5516 04:26:32,788 --> 04:26:35,157 APPLICATIONS ON IN SILICO 5517 04:26:35,157 --> 04:26:37,226 METHODOLOGIES IN DRUG DISCOVERY. 5518 04:26:37,226 --> 04:26:39,828 COME BACK AND TAKE A SHORT BREAK 5519 04:26:39,828 --> 04:26:44,833 AND COME BACK AT 4:15. 5520 04:26:44,833 --> 04:26:46,991 WE'LL START SESSION 3. 5521 04:26:49,240 --> 04:26:50,708 >> WELCOME BACK TO THE LAST 5522 04:26:50,708 --> 04:26:51,642 SESSION OF THE DAY. 5523 04:26:51,642 --> 04:26:55,013 THIS SESSION IS GOING TO BE ON 5524 04:26:55,013 --> 04:27:01,953 APPLICATIONS OF IN SILICO 5525 04:27:01,953 --> 04:27:03,621 METHODOLOGIES AND PLEASED TO 5526 04:27:03,621 --> 04:27:05,590 INTRODUCE DR. RELE PROGRAM 5527 04:27:05,590 --> 04:27:08,359 OFFICER AND SENIOR EXPERT IN 5528 04:27:08,359 --> 04:27:10,495 VACCINE AND THERAPEUTICS PRODUCT 5529 04:27:10,495 --> 04:27:13,798 DEVELOPMENT AND PLATFORM AND 5530 04:27:13,798 --> 04:27:14,899 TECHNOLOGIES AND DEVELOPING 5531 04:27:14,899 --> 04:27:16,834 ADVENTURES AT BARDA. 5532 04:27:16,834 --> 04:27:20,171 HE HAS 23 PLUS YEARS OF 5533 04:27:20,171 --> 04:27:23,574 EXPERIENCE AT THE INTERFACE OF 5534 04:27:23,574 --> 04:27:25,676 BIO TECH MANUFACTURING AND 5535 04:27:25,676 --> 04:27:27,278 OPERATIONS AND NONPROFIT 5536 04:27:27,278 --> 04:27:29,447 ACADEMIA AND MANAGEMENT 5537 04:27:29,447 --> 04:27:30,581 CONSULTING. 5538 04:27:30,581 --> 04:27:40,925 THE FLOOR IS YOURS. 5539 04:27:51,502 --> 04:27:52,770 >> THE HOST HAS STOPPED THE 5540 04:27:52,770 --> 04:27:54,572 VIDEO. 5541 04:27:54,572 --> 04:28:04,916 >> GIVE US A SECOND. 5542 04:28:07,785 --> 04:28:08,052 THERE YOU GO. 5543 04:28:08,052 --> 04:28:08,486 GREAT. 5544 04:28:08,486 --> 04:28:10,054 THE FLOOR IS YOURS. 5545 04:28:10,054 --> 04:28:14,559 >> THANK YOU SO MUCH FOR THE 5546 04:28:14,559 --> 04:28:17,628 KIND INTRODUCTION. 5547 04:28:17,628 --> 04:28:28,072 I'M TALK ABOUT IN SILICO 5548 04:28:31,742 --> 04:28:32,910 METHODOLOGIES IN DRUG DISCOVERY 5549 04:28:32,910 --> 04:28:34,545 AND HOW TO FOR THE WORKSHOP AND 5550 04:28:34,545 --> 04:28:38,583 EXPERTS AND ACADEMIA AND 5551 04:28:38,583 --> 04:28:43,788 INDUSTRY. 5552 04:28:43,788 --> 04:28:46,090 I WANTED TO THANK ALL FOR BEING 5553 04:28:46,090 --> 04:28:51,362 APART OF THIS WORKSHOP. 5554 04:28:51,362 --> 04:28:59,137 LISTENING TO ALL THE TALKS YOU 5555 04:28:59,137 --> 04:29:00,638 WAS DOING THE SYNTHESIS THE OLD 5556 04:29:00,638 --> 04:29:02,507 SCHOOL WAY. 5557 04:29:02,507 --> 04:29:06,577 GREAT TO SEE THE WORK BEING 5558 04:29:06,577 --> 04:29:08,746 DONE. 5559 04:29:08,746 --> 04:29:12,517 RE-EMPHASIZING WHAT'S BEEN SAID, 5560 04:29:12,517 --> 04:29:14,085 BIOLOGY SYSTEMS ARE COMPLEX AND 5561 04:29:14,085 --> 04:29:18,556 ANY TOOLS THAT CAN SIMULATE 5562 04:29:18,556 --> 04:29:24,729 THEIR BEHAVIOR AND THEREFORE 5563 04:29:24,729 --> 04:29:26,564 CONTINUING THE DESIGN AND 5564 04:29:26,564 --> 04:29:29,033 COMPUTATIONAL MODELS AND 5565 04:29:29,033 --> 04:29:35,773 ALGORITHMS AS TOOLS TO PREDICT 5566 04:29:35,773 --> 04:29:42,580 THE MOLECULES AND THE LIGANDS 5567 04:29:42,580 --> 04:29:45,116 AND RECEPTORS PROVIDING INSIGHT 5568 04:29:45,116 --> 04:29:50,555 INTO THE SYNTHESIS AND 5569 04:29:50,555 --> 04:29:53,424 POLYMORPHISM AND PATHWAYS FOR 5570 04:29:53,424 --> 04:29:56,961 NEXT GENERATION TREATMENT 5571 04:29:56,961 --> 04:30:01,165 MODALITIES ARE KEY. 5572 04:30:01,165 --> 04:30:03,534 THEREFORE THE MOLECULAR DESIGN 5573 04:30:03,534 --> 04:30:06,571 AND TESTING AND SCALE UP OF THE 5574 04:30:06,571 --> 04:30:10,241 MOLECULES ARE CRITICAL FOR THE 5575 04:30:10,241 --> 04:30:11,676 DRUG DISCOVERY. 5576 04:30:11,676 --> 04:30:17,582 THIS IS GOING TO BE A 5577 04:30:17,582 --> 04:30:18,549 STIMULATION SESSION ON DRUG 5578 04:30:18,549 --> 04:30:19,450 DISCOVERY. 5579 04:30:19,450 --> 04:30:26,557 LET ME INTRODUCE OUR FIRST 5580 04:30:26,557 --> 04:30:33,898 SPEAKER, SEAN EKINS AND USING 5581 04:30:33,898 --> 04:30:36,834 MACHINE LEARNING AND A.I. 5582 04:30:36,834 --> 04:30:39,937 APPROACHING AFTER HIS POSTDOC AT 5583 04:30:39,937 --> 04:30:42,573 ELI LILLY HE WAS WITH PFIZER AND 5584 04:30:42,573 --> 04:30:48,646 THEN LATER WITH START-UP PHARMA 5585 04:30:48,646 --> 04:30:59,190 AND COMPANIES HE'S ALSO A SOUGHT 5586 04:31:00,658 --> 04:31:02,560 OUT RARE DISEASE ADVISER. 5587 04:31:02,560 --> 04:31:06,063 HIS TOPIC OF PRESENTATION IS 5588 04:31:06,063 --> 04:31:07,298 OUTLINED IN SMALL DRUG 5589 04:31:07,298 --> 04:31:08,966 DISCOVERY. 5590 04:31:08,966 --> 04:31:14,572 THE FLOOR IS YOURS. 5591 04:31:14,572 --> 04:31:17,141 YOU HAVE 25 MINUTES AND IN 22 5592 04:31:17,141 --> 04:31:19,310 MINUTES I'LL INFORM YOU TO WRAP 5593 04:31:19,310 --> 04:31:19,677 UP. 5594 04:31:19,677 --> 04:31:20,945 >> THANK YOU. 5595 04:31:20,945 --> 04:31:25,049 IT'S GREAT TO BE HERE. 5596 04:31:25,049 --> 04:31:27,818 GREAT HONOR TO BE INVITED AND 5597 04:31:27,818 --> 04:31:32,957 WOULD LIKE TO THANK THE 5598 04:31:32,957 --> 04:31:33,357 ORGANIZERS. 5599 04:31:33,357 --> 04:31:36,560 IT'S BEEN STIMULATING LISTENING 5600 04:31:36,560 --> 04:31:38,562 TO THE TALKS. 5601 04:31:38,562 --> 04:31:41,599 I'LL GIVE AN OVERVIEW SINCE 5602 04:31:41,599 --> 04:31:43,134 FINDING THE SMALL COMPANY. 5603 04:31:43,134 --> 04:31:48,272 IT WILL BE QUITE DIFFERENT TO 5604 04:31:48,272 --> 04:31:52,109 MOST THE TALKS WHICH HAVE BEEN 5605 04:31:52,109 --> 04:31:56,714 PRIMARILY ACADEMIC. 5606 04:31:56,714 --> 04:32:00,551 WE'RE A SMALL COMPANY THE FOLKS 5607 04:32:00,551 --> 04:32:02,953 ON THE SLIDE ARE ESSENTIALLY 5608 04:32:02,953 --> 04:32:08,192 RESPONSIBLE FOR THE WORK I'LL 5609 04:32:08,192 --> 04:32:08,426 PRESENT. 5610 04:32:08,426 --> 04:32:09,327 AND WE DEVELOPED VARIOUS 5611 04:32:09,327 --> 04:32:15,199 TECHNOLOGIES AND SOFTWARE FOR 5612 04:32:15,199 --> 04:32:17,935 DRUG DISCOVERY AND APPLICATIONS. 5613 04:32:17,935 --> 04:32:20,004 WHAT'S UNIQUE ABOUT US IS WE 5614 04:32:20,004 --> 04:32:22,373 HAVEN'T BEEN FUNDED BY VENTURE 5615 04:32:22,373 --> 04:32:26,210 CAPITAL OR ANGEL FUNDING AND 5616 04:32:26,210 --> 04:32:27,978 RELIED ON FUNDING FROM NIH AND 5617 04:32:27,978 --> 04:32:29,013 DEPARTMENT OF DEFENSE. 5618 04:32:29,013 --> 04:32:30,314 LOTS OF GRANT WRITING. 5619 04:32:30,314 --> 04:32:34,518 THAT'S ENABLED US TO BRING IN 5620 04:32:34,518 --> 04:32:37,822 ABOUT $23.5 MILLION SO FAR WITH 5621 04:32:37,822 --> 04:32:40,991 COLLABORATORS ON VARIOUS PRO 5622 04:32:40,991 --> 04:32:47,365 PROJECTS SINCE 2016. 5623 04:32:47,365 --> 04:32:49,066 A UNIQUE WAY TO FUND A COMPANY 5624 04:32:49,066 --> 04:32:51,502 AND HAVE LABS ON THE CAMPUS OF 5625 04:32:51,502 --> 04:32:52,570 NC STATE UNIVERSITY. 5626 04:32:52,570 --> 04:32:55,373 WE'RE NOT AFFILIATED WITH THE 5627 04:32:55,373 --> 04:32:56,941 UNIVERSITY BUT TAKE ADVANTAGE OF 5628 04:32:56,941 --> 04:33:01,078 GREAT COLLABORATIONS AND ACCESS 5629 04:33:01,078 --> 04:33:02,413 TO AMAZING STUDENTS AND FACULTY. 5630 04:33:02,413 --> 04:33:05,783 OVER THE YEARS WE'VE BEEN ABLE 5631 04:33:05,783 --> 04:33:10,287 TO PRODUCE A NUMBER OF PATENTS 5632 04:33:10,287 --> 04:33:12,356 AND PUBLISH QUITE WIDELY. 5633 04:33:12,356 --> 04:33:14,959 I CAN'T TALK ABOUT EVERY SINGLE 5634 04:33:14,959 --> 04:33:16,193 THING WE'VE DONE SINCE FOUNDING 5635 04:33:16,193 --> 04:33:18,929 THE COMPANY BUT I WANT TO GIVE 5636 04:33:18,929 --> 04:33:24,235 IDEAS OF WHAT'S POSSIBLE USING 5637 04:33:24,235 --> 04:33:30,574 THE TECHNOLOGIES. 5638 04:33:30,574 --> 04:33:32,977 ALSO POSSIBLY INSPIRE OTHERS TO 5639 04:33:32,977 --> 04:33:34,311 START THEIR OWN COMPANY. 5640 04:33:34,311 --> 04:33:38,549 WE'RE AN STREAMING ORGANIZATION. 5641 04:33:38,549 --> 04:33:40,651 WE BUILD OUR OWN SOFTWARE AND 5642 04:33:40,651 --> 04:33:43,287 LICENSE FOR PREDICTING 5643 04:33:43,287 --> 04:33:45,389 PROPERTIES AND TOXICITY IS A BIG 5644 04:33:45,389 --> 04:33:47,958 ONE I'LL TOUCH ON LATER. 5645 04:33:47,958 --> 04:33:49,794 WE ALSO USE THESE TECHNOLOGIES 5646 04:33:49,794 --> 04:33:52,596 TO HELP DRIVE OUR PIPELINE AND 5647 04:33:52,596 --> 04:33:54,298 FIND MOLECULES AND OPTIMIZE 5648 04:33:54,298 --> 04:33:57,902 MOLECULES AND THIRDLY, WE'RE A 5649 04:33:57,902 --> 04:34:00,237 CONTRACT SERVICES ORGANIZATION. 5650 04:34:00,237 --> 04:34:03,774 PHARMACEUTICAL COMPANIES COME 5651 04:34:03,774 --> 04:34:04,975 WITH US WITH CHALLENGING 5652 04:34:04,975 --> 04:34:06,577 PROBLEMS AND WORK ON SOLVING 5653 04:34:06,577 --> 04:34:08,345 THEM USING OUR TECHNOLOGIES. 5654 04:34:08,345 --> 04:34:16,654 THAT'S SO FAR ENABLED US TO BOOT 5655 04:34:16,654 --> 04:34:18,355 STRAP THE COMPANY AND STARTING 5656 04:34:18,355 --> 04:34:19,990 WITH TRADITIONAL A.I. 5657 04:34:19,990 --> 04:34:20,458 APPROACHES. 5658 04:34:20,458 --> 04:34:22,893 THESE ARE TOOLS YOU'VE SEEN 5659 04:34:22,893 --> 04:34:25,796 QUITE A BIT OF TODAY ALREADY 5660 04:34:25,796 --> 04:34:28,299 TOOLS THAT TAKE THE DATA SET 5661 04:34:28,299 --> 04:34:31,235 WHETHER BY ACTIVITY OR PHYSICAL 5662 04:34:31,235 --> 04:34:32,102 CHEMICAL PROPERTY FOR EXAMPLE 5663 04:34:32,102 --> 04:34:34,505 AND ALLOW YOU TO BUILD A 5664 04:34:34,505 --> 04:34:35,306 PREDICTIVE MODEL AND THEN 5665 04:34:35,306 --> 04:34:38,576 OBVIOUSLY GENERATE PREDICTIONS 5666 04:34:38,576 --> 04:34:42,613 FOR MOLECULES WHETHER FROM A 5667 04:34:42,613 --> 04:34:44,615 VIRTUAL LIBRARY OR AVAILABLE 5668 04:34:44,615 --> 04:34:48,252 LIBRARY OF MOLECULES. 5669 04:34:48,252 --> 04:34:49,720 THEY DON'T NECESSARILY ALLOW TO 5670 04:34:49,720 --> 04:34:51,622 YOU CREATE ANYTHING NEW BUT MINE 5671 04:34:51,622 --> 04:34:53,691 THE DATA AND FIND NEW MOLECULES 5672 04:34:53,691 --> 04:34:54,391 WITH NEW USES. 5673 04:34:54,391 --> 04:34:56,093 I'LL TRY TO SHOW SOME OF THOSE 5674 04:34:56,093 --> 04:35:04,435 EXAMPLES AFTER THE SLIDE. 5675 04:35:04,435 --> 04:35:09,874 THE SECOND APPROACH IS MORE 5676 04:35:09,874 --> 04:35:14,578 GENERATIVE A.I. WHERE YOU CAN 5677 04:35:14,578 --> 04:35:17,815 LARGE LANGUAGE MODELS OR 5678 04:35:17,815 --> 04:35:23,988 APPROACHES AND ENABLE MAKING 5679 04:35:23,988 --> 04:35:29,527 DECISIONS GOING BEYOND THE DATA 5680 04:35:29,527 --> 04:35:32,963 TO PROPOSE NEW NOVEL MOLECULES 5681 04:35:32,963 --> 04:35:36,600 AND AS A SMALL COMPANY WE'RE AT 5682 04:35:36,600 --> 04:35:38,769 THE CUTTING EDGE OF USING THE 5683 04:35:38,769 --> 04:35:39,203 TOOLS. 5684 04:35:39,203 --> 04:35:42,306 WHEN WE SEE WHAT BIG PHARMA ARE 5685 04:35:42,306 --> 04:35:44,975 DOING WITH WAY MORE INVESTMENT 5686 04:35:44,975 --> 04:35:48,045 THEN THE $23.5 MILLION WE 5687 04:35:48,045 --> 04:35:49,947 BROUGHT IN FROM NIH AND DOD IT 5688 04:35:49,947 --> 04:35:53,350 MAKES YOU REALIZE BIG PHARMA HAS 5689 04:35:53,350 --> 04:35:56,954 MASSIVELY UNDER INVESTED IN A.I. 5690 04:35:56,954 --> 04:36:01,492 TO DATE. 5691 04:36:01,492 --> 04:36:04,094 TO SOME PEOPLE IT MAY SEEM 5692 04:36:04,094 --> 04:36:05,596 CONTROVERSIAL BUT MANY HAVE TOLD 5693 04:36:05,596 --> 04:36:08,566 US THEY'RE A DECADE-PLUS BEHIND. 5694 04:36:08,566 --> 04:36:10,267 THERE'S STILL A LOT OF 5695 04:36:10,267 --> 04:36:11,635 OPPORTUNITY FOR THESE TYPES OF 5696 04:36:11,635 --> 04:36:13,704 COMPANIES TO INVEST IN THIS 5697 04:36:13,704 --> 04:36:13,971 SPACE. 5698 04:36:13,971 --> 04:36:18,576 THE TYPES OF TOOLS WE INITIALLY 5699 04:36:18,576 --> 04:36:21,579 USED STARTED WITH BAYESIAN 5700 04:36:21,579 --> 04:36:24,214 ALGORITHMS AND RANDOM FOREST, 5701 04:36:24,214 --> 04:36:24,481 ETCETERA. 5702 04:36:24,481 --> 04:36:26,317 OVER OF THE YEARS WE EXPANDED TO 5703 04:36:26,317 --> 04:36:27,551 LOOK AT MANY ALGORITHMS IN 5704 04:36:27,551 --> 04:36:32,957 PARALLEL THE CLASSIFICATION AND 5705 04:36:32,957 --> 04:36:38,362 REGRESSION AND MORE 5706 04:36:38,362 --> 04:36:39,997 SOPHISTICATED TOOLS LIKE 5707 04:36:39,997 --> 04:36:41,098 DIFFERENT APPROACHES. 5708 04:36:41,098 --> 04:36:44,602 A WHOLE MIX OF ALGORITHMS. 5709 04:36:44,602 --> 04:36:46,070 AS I'LL TOUCH ON LATER, PICKING 5710 04:36:46,070 --> 04:36:48,739 WHICH ALGORITHMS TO USE FOR A 5711 04:36:48,739 --> 04:36:50,240 PARTICULAR DATA SET IS IN ITSELF 5712 04:36:50,240 --> 04:36:56,947 A CHALLENGE. 5713 04:36:56,947 --> 04:36:59,817 INITIALLY, WE WERE TRYING TO 5714 04:36:59,817 --> 04:37:03,754 FOCUS ON MOLECULES AND TRY TO 5715 04:37:03,754 --> 04:37:06,557 DESIGN, BUY, TEST, ANALYZE. 5716 04:37:06,557 --> 04:37:08,258 WE WEREN'T THINKING OF MAKING 5717 04:37:08,258 --> 04:37:08,959 MOLECULES BECAUSE THEY WERE 5718 04:37:08,959 --> 04:37:11,562 BEYOND WHAT WE WERE ABLE TO DO. 5719 04:37:11,562 --> 04:37:12,496 INITIALLY AS A SMALL COMPANY WE 5720 04:37:12,496 --> 04:37:14,231 DIDN'T HAVE A LAB. 5721 04:37:14,231 --> 04:37:16,333 WE WERE USING TECHNOLOGIES AN 5722 04:37:16,333 --> 04:37:17,968 BUILDING TECHNOLOGIES SO WE 5723 04:37:17,968 --> 04:37:20,971 COULD SCREEN THROUGH FDA 5724 04:37:20,971 --> 04:37:25,709 APPROVED DRUGS, ETCETERA AND 5725 04:37:25,709 --> 04:37:27,678 SEND THEM OFF TO COLLABORATORS. 5726 04:37:27,678 --> 04:37:29,913 ONE TARGET AT A TIME OR ONE 5727 04:37:29,913 --> 04:37:37,488 PHENOTYPIC ASSAY AT A TIME. 5728 04:37:37,488 --> 04:37:39,957 THIS LED US TO PUT TOGETHER TEST 5729 04:37:39,957 --> 04:37:43,560 CASES TO SAY LET'S DEMONSTRATE 5730 04:37:43,560 --> 04:37:50,401 THIS AND PRODUCE SOFTWARE AND 5731 04:37:50,401 --> 04:38:00,911 BUNDLE DIFFERENT ALGORITHMS AND 5732 04:38:00,911 --> 04:38:05,315 USE THEM FOR PREDICTIONS. 5733 04:38:05,315 --> 04:38:07,651 WHAT YOU'LL SEE ARE THINGS WE'VE 5734 04:38:07,651 --> 04:38:10,120 DONE DEVELOPING THE 5735 04:38:10,120 --> 04:38:10,454 TECHNOLOGIES. 5736 04:38:10,454 --> 04:38:13,023 AND THEY OBVIOUSLY LED US TO 5737 04:38:13,023 --> 04:38:14,358 DEVELOP TOOLS THAT WE HAVE NOW 5738 04:38:14,358 --> 04:38:14,858 BEEN ABLE TO LICENSE TO 5739 04:38:14,858 --> 04:38:19,663 COMPANIES. 5740 04:38:19,663 --> 04:38:22,332 ONE OF THE EARLY EXAMPLES WE 5741 04:38:22,332 --> 04:38:26,103 PUBLISHED BACK IN 2015 WAS 5742 04:38:26,103 --> 04:38:30,441 AROUND THE EBOLA OUTBREAK OF 5743 04:38:30,441 --> 04:38:30,641 2014. 5744 04:38:30,641 --> 04:38:34,945 NOWADAYS WE THINK OF OUTBREAKS 5745 04:38:34,945 --> 04:38:39,216 LIKE SARS COVID AND THERE WAS A 5746 04:38:39,216 --> 04:38:40,150 BIG EBOLA OUTBREAK. 5747 04:38:40,150 --> 04:38:44,388 I WAS INTERESTED IN TRYING TO 5748 04:38:44,388 --> 04:38:46,190 ACCESS BUILD MACHINE LEARNING 5749 04:38:46,190 --> 04:38:49,093 MODEL TO TRY TO IDENTIFY NEW 5750 04:38:49,093 --> 04:38:50,561 ANTIVIRALS. 5751 04:38:50,561 --> 04:38:58,569 A COLLABORATOR AT THE TIME BUILD 5752 04:38:58,569 --> 04:39:02,139 A MODEL WE WERE ABLE TO USE TO 5753 04:39:02,139 --> 04:39:03,574 VIRTUALLY SCREEN A LIBRARY OF 5754 04:39:03,574 --> 04:39:06,243 ABOUT 2500 MOLECULES IN THE 5755 04:39:06,243 --> 04:39:06,543 LIBRARY. 5756 04:39:06,543 --> 04:39:11,548 THAT THEN ENABLED US TO RANK THE 5757 04:39:11,548 --> 04:39:14,551 MOLECULE AND SELECTED THREE OF 5758 04:39:14,551 --> 04:39:14,985 THEM. 5759 04:39:14,985 --> 04:39:17,387 TO TEST WITH A COLLABORATOR IN 5760 04:39:17,387 --> 04:39:18,188 VITRO. 5761 04:39:18,188 --> 04:39:20,691 WE WERE ABLE TO FIND ALL THREE 5762 04:39:20,691 --> 04:39:22,693 MOLECULES THAT HAVEN'T BEEN 5763 04:39:22,693 --> 04:39:26,029 TESTED PREVIOUSLY WERE NANOMOLE 5764 04:39:26,029 --> 04:39:27,364 ACTIVES IN THIS IN VITRO ASSAY. 5765 04:39:27,364 --> 04:39:30,567 THAT WAS SURPRISING FOR A NUMBER 5766 04:39:30,567 --> 04:39:38,542 OF REASONS. 5767 04:39:38,542 --> 04:39:43,313 ONE WAS AN ANTI-MARLARIAL 5768 04:39:43,313 --> 04:39:46,550 APPROVED SEVERAL YEARS AFTER THE 5769 04:39:46,550 --> 04:39:46,884 PUBLICATION. 5770 04:39:46,884 --> 04:39:50,888 AND WE WERE ABLE TO USE THE ONE 5771 04:39:50,888 --> 04:39:53,657 GRAPH TO GET AN R21 FROM NCATS. 5772 04:39:53,657 --> 04:39:56,460 SO WE'RE VERY GRATEFUL TO THEM. 5773 04:39:56,460 --> 04:39:58,262 THAT KICKED OFF WORK WITHIN OUR 5774 04:39:58,262 --> 04:39:59,830 COMPANY IN TERMS OF EVALUATING 5775 04:39:59,830 --> 04:40:03,200 THE MOLECULE NOT ONLY FOR EBOLA 5776 04:40:03,200 --> 04:40:08,071 BUT ALSO AGAINST OTHER VIRUSES. 5777 04:40:08,071 --> 04:40:09,406 OVER SUBSEQUENT YEARS WE'VE BEEN 5778 04:40:09,406 --> 04:40:15,012 ABLE TO USE THIS MOLECULE 5779 04:40:15,012 --> 04:40:15,579 IDENTIFIED THROUGH MACHINE 5780 04:40:15,579 --> 04:40:18,215 LEARNING AND CHARACTERIZE IT AND 5781 04:40:18,215 --> 04:40:21,251 SHOWED IT ACCUMULATES IN THE 5782 04:40:21,251 --> 04:40:26,990 LYSOSOMES AND CHANGING PH AND AN 5783 04:40:26,990 --> 04:40:31,728 INHIBITOR AND MOLAR AND 5784 04:40:31,728 --> 04:40:35,766 PERFORMED DOCKI IING TO EVALUA 5785 04:40:35,766 --> 04:40:39,002 TARGETS AND FOLLOWED THIS UP 5786 04:40:39,002 --> 04:40:42,372 WITH DOING BINDING AND SHOWED 5787 04:40:42,372 --> 04:40:47,010 IT'S A PRETTY POTENT LOW MICRO 5788 04:40:47,010 --> 04:40:50,314 MOLAR BINDER TO THE TARGET AND 5789 04:40:50,314 --> 04:40:51,949 LOOKED AT HOW IT MAY INTERFERE 5790 04:40:51,949 --> 04:40:55,853 WITH ACCESS OF THE VIRUS INTO 5791 04:40:55,853 --> 04:40:56,053 CELLS. 5792 04:40:56,053 --> 04:41:00,691 WE SHOWED WITH A PSEUDO VIRUS 5793 04:41:00,691 --> 04:41:03,393 THE MOLECULE WAS ABLE TO BLOCK 5794 04:41:03,393 --> 04:41:03,760 UPTAKE. 5795 04:41:03,760 --> 04:41:05,462 SO REALLY INTERESTING OVER THE 5796 04:41:05,462 --> 04:41:07,631 YEARS WE'VE BEEN ABLE TO FILL 5797 04:41:07,631 --> 04:41:11,568 OUT WHAT'S UNDERSTOOD ABOUT THE 5798 04:41:11,568 --> 04:41:13,303 MOLECULE TRADITIONALLY THOUGHT 5799 04:41:13,303 --> 04:41:23,914 OF AS AN ANTI-MALARIAL AND HAVE 5800 04:41:34,057 --> 04:41:35,826 COULD HAVE PK INFORMING. 5801 04:41:35,826 --> 04:41:38,595 THIS ENABLES US TO DO A SINGLE 5802 04:41:38,595 --> 04:41:46,904 DOSE STUDY IN MICE AND COULD DO 5803 04:41:46,904 --> 04:41:48,438 A TIME ASSAY AS WELL. 5804 04:41:48,438 --> 04:41:50,507 PROVIDING THE DRUG WINDOWS AFTER 5805 04:41:50,507 --> 04:41:51,675 DOSING WITH THE EBOLA VIRUS AND 5806 04:41:51,675 --> 04:41:53,944 SHOWING WE COULD HAVE PROTECTION 5807 04:41:53,944 --> 04:41:54,578 UP TO 24 HOURS. 5808 04:41:54,578 --> 04:42:00,484 SO WE'RE ABLE TO LOOK AT PLAQUE 5809 04:42:00,484 --> 04:42:01,952 ASSAY TO ASSESS THE VIRAL LEVELS 5810 04:42:01,952 --> 04:42:06,089 AND THIS WORK WAS PUBLISHED BY 5811 04:42:06,089 --> 04:42:08,959 OUR LAB AND HE'S WITHIN 5812 04:42:08,959 --> 04:42:10,193 RESPONSIBLE FOR MOST THE WORK 5813 04:42:10,193 --> 04:42:13,297 DURING THIS PERIOD AND SINCE. 5814 04:42:13,297 --> 04:42:15,866 AND ULTIMATELY THIS ENABLED US 5815 04:42:15,866 --> 04:42:18,568 TO OBTAIN AN ORPHAN DRUG 5816 04:42:18,568 --> 04:42:20,570 DESIGNATION FROM THE FDA IN 5817 04:42:20,570 --> 04:42:21,004 2019. 5818 04:42:21,004 --> 04:42:22,639 FOLLOWING THIS WE WERE ABLE TO 5819 04:42:22,639 --> 04:42:25,008 CHARACTERIZE THE MOLECULE AND 5820 04:42:25,008 --> 04:42:27,911 SHOW WE COULD DOSE AT A SINGLE 5821 04:42:27,911 --> 04:42:31,381 DOSE AND WE HAD SOME PROTECTION 5822 04:42:31,381 --> 04:42:32,983 AGAINST THE EBOLA VIRUS. 5823 04:42:32,983 --> 04:42:35,018 IT WASN'T AS CONVINCING AS IN 5824 04:42:35,018 --> 04:42:37,654 THE MOUSE AND PRIMARILY DUE TO 5825 04:42:37,654 --> 04:42:40,590 THE INCREASED METABOLISM IN THE 5826 04:42:40,590 --> 04:42:41,725 GUINEA PIG. 5827 04:42:41,725 --> 04:42:48,565 ADDITIONAL STUDIES WITH 5828 04:42:48,565 --> 04:42:50,567 COLLABORATORS FOUND IT WAS 5829 04:42:50,567 --> 04:42:52,769 EFFECTIVE AGAINST THREE STRAINS. 5830 04:42:52,769 --> 04:42:57,007 AND THESE ASSAYS WERE ABLE TO 5831 04:42:57,007 --> 04:43:02,479 SHOW A SINGLE MICRO MOLAR. 5832 04:43:02,479 --> 04:43:04,147 THE NEXT STAGE IS TO GET INTO 5833 04:43:04,147 --> 04:43:07,351 NON HUMAN PRIMATES AND WE HAVE 5834 04:43:07,351 --> 04:43:08,986 TO OBTAIN SEVERAL MILLIONS OF 5835 04:43:08,986 --> 04:43:10,587 FUNDING AND WE'RE CURRENTLY 5836 04:43:10,587 --> 04:43:11,355 PUTTING GRANTS IN TO GET THAT 5837 04:43:11,355 --> 04:43:14,558 WORK DONE. 5838 04:43:14,558 --> 04:43:17,461 WHAT WAS INTERESTING FOR US 5839 04:43:17,461 --> 04:43:19,963 AROUND 2020, WE COULD NOT HAVE 5840 04:43:19,963 --> 04:43:23,700 PREDICTED THERE WOULD BE A MAJOR 5841 04:43:23,700 --> 04:43:25,969 OUTBREAK OF A NEW VIRUS THAT 5842 04:43:25,969 --> 04:43:30,540 ULTIMATELY BECAME SARS COV2. 5843 04:43:30,540 --> 04:43:34,044 WE WERE EXCITED BECAUSE 5844 04:43:34,044 --> 04:43:35,879 REMDESIVIR ANOTHER EBOLA DRUG 5845 04:43:35,879 --> 04:43:38,749 SHOWED TO BE EFFICACIOUS AND 5846 04:43:38,749 --> 04:43:40,484 WANTED TO LOOK AT THE ACTIVITY. 5847 04:43:40,484 --> 04:43:42,486 WE COLLABORATED WITH A NUMBER OF 5848 04:43:42,486 --> 04:43:44,755 GROUPS IN THE U.S. AND OUTSIDE 5849 04:43:44,755 --> 04:43:46,890 TO THE U.S. TO TRY TO TEST THE 5850 04:43:46,890 --> 04:43:47,691 MOLECULE. 5851 04:43:47,691 --> 04:43:52,863 INITIALLY WE WERE ABLE TO SHOW 5852 04:43:52,863 --> 04:43:59,569 ACTIVITY IN CELL LINES WE FOUND 5853 04:43:59,569 --> 04:44:01,004 ACTIVITY IN THE CELLS AND THEN 5854 04:44:01,004 --> 04:44:02,406 WE WERE ABLE TO TEST THE 5855 04:44:02,406 --> 04:44:07,778 MOLECULE IN A MOUSE MODEL WITH A 5856 04:44:07,778 --> 04:44:10,514 GROUP IN BRAZIL AND SHOW IT'S 5857 04:44:10,514 --> 04:44:13,550 EFFICACIOUS AND IT HAD AN IMPACT 5858 04:44:13,550 --> 04:44:15,719 ON INFLAMMATORY CYTOKINES AND 5859 04:44:15,719 --> 04:44:18,555 WAS AN ANTI-INFLAMMATORY 5860 04:44:18,555 --> 04:44:20,857 MOLECULE AND DECREASED CELL 5861 04:44:20,857 --> 04:44:26,563 INFILTRATION AND SHOWED IT WAS 5862 04:44:26,563 --> 04:44:28,932 AN INHIBITOR IN VITRO. 5863 04:44:28,932 --> 04:44:30,567 THE MOLECULE APPEARS TO TARGET 5864 04:44:30,567 --> 04:44:32,135 DIFFERENT TARGETS IN DIFFERENT 5865 04:44:32,135 --> 04:44:37,908 VIRUSES BASED ON THE DATA WE 5866 04:44:37,908 --> 04:44:38,108 HAVE. 5867 04:44:38,108 --> 04:44:41,912 AND THE PHARMACEUTICAL COMPANY 5868 04:44:41,912 --> 04:44:43,280 TOOK THIS COMBINATION WHICH THEY 5869 04:44:43,280 --> 04:44:46,950 MADE FOR MALARIA AND TOOK IT 5870 04:44:46,950 --> 04:44:48,251 THROUGH PHASE 3 CLINICAL TRIAL 5871 04:44:48,251 --> 04:44:52,456 AS WELL IN HUMANS. 5872 04:44:52,456 --> 04:44:55,826 AT ABOUT THE SAME TIME WE WERE 5873 04:44:55,826 --> 04:44:59,029 WORKING ON THAT WE HAD DONE 5874 04:44:59,029 --> 04:45:00,564 VIRTUAL SCREEN AGAIN WITH THE 5875 04:45:00,564 --> 04:45:06,002 DATA SET THIS TIME FOR PARASITE 5876 04:45:06,002 --> 04:45:07,504 RESULTING IN A DISEASE. 5877 04:45:07,504 --> 04:45:18,048 THIS IS WITH COLLABORATORS UCSD 5878 04:45:21,084 --> 04:45:25,422 AND IN THE STUDY WE LOOKED AT 5879 04:45:25,422 --> 04:45:28,358 BUILDING A MODEL WITH 4,000 5880 04:45:28,358 --> 04:45:29,359 MODELS FROM THE LITERATURE AND 5881 04:45:29,359 --> 04:45:32,295 THEN SCREENED THE LIBRARY OF 5882 04:45:32,295 --> 04:45:40,537 ABOUT 7,500 MOLECULES AND PICKED 5883 04:45:40,537 --> 04:45:45,642 FOR SOME FOR TESTING AND THRN WE 5884 04:45:45,642 --> 04:45:46,776 WERE ABLE TO TEST IN AP MALLS 5885 04:45:46,776 --> 04:45:50,981 AND IN THE ACUTE MODEL AND 5886 04:45:50,981 --> 04:45:53,650 SHOWED IT HAD GOOD EFFICACY AND 5887 04:45:53,650 --> 04:45:56,520 AGAIN WE USED THIS IN VIVO DATA 5888 04:45:56,520 --> 04:45:58,221 TO OBTAIN THE ORPHAN DRUG 5889 04:45:58,221 --> 04:45:58,922 DESIGNATION FOR THIS PARTICULAR 5890 04:45:58,922 --> 04:46:03,493 USE. 5891 04:46:03,493 --> 04:46:06,563 THAT ALSO LED US TO OBTAIN 5892 04:46:06,563 --> 04:46:10,567 ANOTHER GRANT THE UG2, UG3 GRANT 5893 04:46:10,567 --> 04:46:13,103 THAT FUNDED ANOTHER GROUP TO DO 5894 04:46:13,103 --> 04:46:15,005 MORE IN VIVO WORK AND PK. 5895 04:46:15,005 --> 04:46:18,575 IN THIS CASE WE WERE LOOKING AT 5896 04:46:18,575 --> 04:46:20,343 DIFFERENT STRAINS AND TRYING TO 5897 04:46:20,343 --> 04:46:23,413 SHOW WHETHER THERE WAS EFFICACY 5898 04:46:23,413 --> 04:46:28,385 IN THE CHRONIC MODEL OF THE 5899 04:46:28,385 --> 04:46:28,718 DISEASE. 5900 04:46:28,718 --> 04:46:31,721 WHAT WE FOUND USING FROM IP 5901 04:46:31,721 --> 04:46:33,623 DOSING TO DOSING THE DRUG 5902 04:46:33,623 --> 04:46:35,425 ORALLY, WE'RE ABLE TO SHOW 5903 04:46:35,425 --> 04:46:37,928 EFFICACY STILL IN ACUTE MODELS 5904 04:46:37,928 --> 04:46:40,797 WITH A DIFFERENT STRAIN OF THE 5905 04:46:40,797 --> 04:46:41,298 PARASITE. 5906 04:46:41,298 --> 04:46:41,932 UNFORTUNATELY WE WEREN'T ABLE TO 5907 04:46:41,932 --> 04:46:44,668 SHOW EFFICACY IN A LONGER 5908 04:46:44,668 --> 04:46:46,570 CHRONIC MODEL THAT GOES ON FOR 5909 04:46:46,570 --> 04:46:48,305 SIX MONTHS OR SO. 5910 04:46:48,305 --> 04:46:49,673 SO WE THINK THERE'S STILL SOME 5911 04:46:49,673 --> 04:46:52,142 WAY TO GO IN TERMS OF OPTIMIZING 5912 04:46:52,142 --> 04:46:55,478 THE MODELS HERE IN TERMS OF THE 5913 04:46:55,478 --> 04:46:57,013 STRAINS. 5914 04:46:57,013 --> 04:47:01,418 WE MIGHT HAVE A STRAIN DEPENDENT 5915 04:47:01,418 --> 04:47:02,319 EFFECT. 5916 04:47:02,319 --> 04:47:03,386 INTERESTINGLY THE HIGH LEVEL 5917 04:47:03,386 --> 04:47:06,556 VIEW GOING FROM MACHINE LEARNING 5918 04:47:06,556 --> 04:47:11,628 TO IDENTIFYING MOLECULES THROUGH 5919 04:47:11,628 --> 04:47:13,797 IN VIVO WE'VE DONE SCREENS FOR 5920 04:47:13,797 --> 04:47:15,332 OTHER TARGETS AND DONE WORK AS 5921 04:47:15,332 --> 04:47:16,666 HIGHLIGHTED BRIEFLY ON THIS 5922 04:47:16,666 --> 04:47:18,868 SLIDE WITH A COUPLE PROVOCATIONS 5923 04:47:18,868 --> 04:47:25,642 AGAIN THIS WORK WAS LED BY ANNA 5924 04:47:25,642 --> 04:47:28,478 POOL IN OUR GROUP AND WORKING 5925 04:47:28,478 --> 04:47:32,582 WITH KEN JACOBSON AND IT 5926 04:47:32,582 --> 04:47:37,754 DEMONSTRATES FINDING NEW USES 5927 04:47:37,754 --> 04:47:39,656 FOR FDA APPROVED DRUGS AND THEY 5928 04:47:39,656 --> 04:47:45,028 HAVE ACTIVITY IN BINDING S FOR A 5929 04:47:45,028 --> 04:47:48,031 NUMBER OF ADENOSINE RECEPTORS 5930 04:47:48,031 --> 04:47:50,567 AND HAVING ACTIVITY IN IN TAR 5931 04:47:50,567 --> 04:47:54,904 GETS AS WELL. 5932 04:47:54,904 --> 04:47:57,140 WE TRIED THE TECHNOLOGIES IN 5933 04:47:57,140 --> 04:47:59,409 TERMS OF PHENOTYPIC SCREENS AND 5934 04:47:59,409 --> 04:48:01,111 MER TARGET BASED SCREENS. 5935 04:48:01,111 --> 04:48:03,046 THEN A MORE RECENT EXAMPLE WAS 5936 04:48:03,046 --> 04:48:06,549 AT THE REQUEST OF AN ACADEMIC AT 5937 04:48:06,549 --> 04:48:09,886 THE UNIVERSITY OF KENTUCKY WHO 5938 04:48:09,886 --> 04:48:11,454 IDENTIFIED A COUPLE CHEMOKINE 5939 04:48:11,454 --> 04:48:14,324 RECEPTORS AS BEING INVOLVED IN 5940 04:48:14,324 --> 04:48:14,557 STROKE. 5941 04:48:14,557 --> 04:48:17,127 HE WONDERED IF WE COULD FIND AN 5942 04:48:17,127 --> 04:48:21,031 FDA APPROVED DRUG THAT MAY BE AN 5943 04:48:21,031 --> 04:48:25,835 ANTAGONIST AND CURATED A COUPLE 5944 04:48:25,835 --> 04:48:27,037 DATA SETS AND BUILT MACHINE 5945 04:48:27,037 --> 04:48:28,838 LEARNING MODELS AND USED THOSE 5946 04:48:28,838 --> 04:48:30,507 TO SCREEN A LIBRARY OF FDA 5947 04:48:30,507 --> 04:48:31,574 APPROVED DRUGS. 5948 04:48:31,574 --> 04:48:37,847 WE'RE ABLE TO FIND A NUMBER OF 5949 04:48:37,847 --> 04:48:38,581 KINASE I 5950 04:48:38,581 --> 04:48:42,585 KIN 5951 04:48:42,585 --> 04:48:45,055 KIN 5952 04:48:45,055 --> 04:48:48,525 KINASE INHIBITORS AND ONE HAD A 5953 04:48:48,525 --> 04:48:50,126 TARGET FOR HIV. 5954 04:48:50,126 --> 04:48:51,728 SO WE'VE DONE PSEUDO VIRUS WORK 5955 04:48:51,728 --> 04:48:53,563 TO SHOW THE MOLECULE ACTUALLY 5956 04:48:53,563 --> 04:48:55,265 BLOCKS ENTRY AS WELL. 5957 04:48:55,265 --> 04:48:57,300 SO REALLY INTERESTING HOW WE'VE 5958 04:48:57,300 --> 04:48:59,669 BEEN ABLE TO TAKE THIS BASIC 5959 04:48:59,669 --> 04:49:01,204 MACHINE LEARNING APPROACH AND GO 5960 04:49:01,204 --> 04:49:06,142 AFTER DIFFERENT TYPES OF 5961 04:49:06,142 --> 04:49:06,376 TARGETS. 5962 04:49:06,376 --> 04:49:10,080 WE'VE ALSO LOOKED AT ENZYMES AND 5963 04:49:10,080 --> 04:49:12,248 LOOKING AT AN IMPORTANT ENZYME 5964 04:49:12,248 --> 04:49:19,789 FOR A NUMBER OF CNS DISEASES AND 5965 04:49:19,789 --> 04:49:21,725 TARGET THROUGH MOLECULE AND 5966 04:49:21,725 --> 04:49:23,693 ENVIRONMENT, PESTICIDES, 5967 04:49:23,693 --> 04:49:24,227 ETCETERA. 5968 04:49:24,227 --> 04:49:28,264 AND THERE ISN'T CURRENTLY A 5969 04:49:28,264 --> 04:49:30,533 SELECTIVE INHIBITOR THAT'S FDA 5970 04:49:30,533 --> 04:49:31,401 APPROVED. 5971 04:49:31,401 --> 04:49:32,302 SO WE LOOKED AT DIFFERENT 5972 04:49:32,302 --> 04:49:33,036 MACHINE LEARNING APPROACHES TO 5973 04:49:33,036 --> 04:49:38,408 TRY TO COME UP ONE SELECTIVE 5974 04:49:38,408 --> 04:49:38,675 MOLECULE. 5975 04:49:38,675 --> 04:49:42,545 SO WE BUILT MODELS WITH A COUPLE 5976 04:49:42,545 --> 04:49:47,984 THOUSAND MOLECULE AND CURATED 5977 04:49:47,984 --> 04:49:52,922 SOME LITERATURE WE CAME UP WITH 5978 04:49:52,922 --> 04:49:53,590 MACHINE LEARNING APPROACHES AND 5979 04:49:53,590 --> 04:49:56,359 DEEP LEARNING AND A CONTRASTED 5980 04:49:56,359 --> 04:49:57,727 LEARNING THAT WILL HOPEFULLY 5981 04:49:57,727 --> 04:49:59,662 ENABLE US TO DO SELECTIVE 5982 04:49:59,662 --> 04:49:59,963 MOLECULE. 5983 04:49:59,963 --> 04:50:04,567 ONCE WE VALIDATED THE MODELS 5984 04:50:04,567 --> 04:50:06,770 THIS IS WORK WITH A POSTDOC IN 5985 04:50:06,770 --> 04:50:11,441 OUR GROUP AND WE DID A SCREEN OF 5986 04:50:11,441 --> 04:50:14,144 ABOUT 5 MILLION COMMERCIAL 5987 04:50:14,144 --> 04:50:16,613 MOLECULES OR TESTING IN VITRO 5988 04:50:16,613 --> 04:50:20,550 AND THEN WE PURCHASED ABOUT 20 5989 04:50:20,550 --> 04:50:22,552 MOLECULES FOR TESTING AND 5990 04:50:22,552 --> 04:50:24,087 FOLLOWED UP WITH DOSE RESPONSE 5991 04:50:24,087 --> 04:50:27,090 DATA AND WERE ABLE TO GET A 35% 5992 04:50:27,090 --> 04:50:33,062 HIT RATE AND ONE MOLECULE HAD AN 5993 04:50:33,062 --> 04:50:37,534 IC50 AND WAS COMPARABLE TO 5994 04:50:37,534 --> 04:50:40,170 CONTROL MOLECULE OF ABOUT HALF A 5995 04:50:40,170 --> 04:50:40,370 MOLAR. 5996 04:50:40,370 --> 04:50:41,571 FASCINATING APPROACH TRYING TO 5997 04:50:41,571 --> 04:50:42,071 FIND MOLECULES THAT ARE 5998 04:50:42,071 --> 04:50:45,175 SELECTIVE. 5999 04:50:45,175 --> 04:50:48,578 ALL THESE ARE GOING AFTER TRYING 6000 04:50:48,578 --> 04:50:51,114 TO FIND BIO OPTIC MOLECULES. 6001 04:50:51,114 --> 04:50:53,016 THESE ARE FRAET EXAMPLES BUT 6002 04:50:53,016 --> 04:50:54,551 THERE'S A GROWING AMOUNT OF DATA 6003 04:50:54,551 --> 04:50:57,420 TRYING TO UNDERSTAND THE 6004 04:50:57,420 --> 04:50:59,289 PROPERTIES OF MOLECULES AND 6005 04:50:59,289 --> 04:51:01,458 THOUGH THESE DATA SETS ARE 6006 04:51:01,458 --> 04:51:03,426 NOWHERE NEAR THE AMOUNT A LARGE 6007 04:51:03,426 --> 04:51:05,395 PHARMA WOULD HAVE, THEY'RE 6008 04:51:05,395 --> 04:51:08,965 ENOUGH INFORMATION TO BUILD 6009 04:51:08,965 --> 04:51:09,566 CONVINCING MACHINE LEARNING 6010 04:51:09,566 --> 04:51:11,401 MODEL TO TAKE ADVANTAGE OF. 6011 04:51:11,401 --> 04:51:14,838 I'LL SHOW ONE EXAMPLE OF RECENT 6012 04:51:14,838 --> 04:51:16,473 PUBLICATION FROM LAST YEAR WHERE 6013 04:51:16,473 --> 04:51:19,642 WE WERE CURATING FOR NOT ONLY 6014 04:51:19,642 --> 04:51:23,279 RATS BUT ALSO MOUSE AND VARIOUS 6015 04:51:23,279 --> 04:51:25,982 FISH SPECIES IN ORDER TO BUILD 6016 04:51:25,982 --> 04:51:28,485 CLASSIFICATION AND REGRESSION 6017 04:51:28,485 --> 04:51:33,056 MODELS TO PREDICT LC50 AND THESE 6018 04:51:33,056 --> 04:51:38,228 WE CAN COMBINE AND MAKE 6019 04:51:38,228 --> 04:51:40,597 AVAILABLE IN A MEGA TOX AND 6020 04:51:40,597 --> 04:51:41,397 PULLED TOGETHER A COLLECTION 6021 04:51:41,397 --> 04:51:46,970 MUCH -- OF MODELS AND ONE 6022 04:51:46,970 --> 04:51:50,573 CHALLENGE IS TO COME UP WITH A 6023 04:51:50,573 --> 04:51:53,076 WAY TO HONE IN ON WHICH MODEL TO 6024 04:51:53,076 --> 04:51:54,010 USE WITH WHICH PARTICULAR DATA 6025 04:51:54,010 --> 04:51:54,310 SET. 6026 04:51:54,310 --> 04:51:57,046 WE LOOKED AT THIS AND A RECENT 6027 04:51:57,046 --> 04:51:58,548 PUBLICATION GOES INTO THIS GOLD 6028 04:51:58,548 --> 04:52:00,183 THE GOLDILOCKS. 6029 04:52:00,183 --> 04:52:02,285 I RECOMMEND CHECKING THAT OUT. 6030 04:52:02,285 --> 04:52:05,288 AND THE LAST EXAMPLE I'D LAKE TO 6031 04:52:05,288 --> 04:52:08,024 PROVIDE IS OUR WORK AROUND USING 6032 04:52:08,024 --> 04:52:08,958 GENERATIVE A.I. 6033 04:52:08,958 --> 04:52:10,593 WE'VE BUILT ON SOME TOOLS THAT 6034 04:52:10,593 --> 04:52:14,264 ARE AVAILABLE. 6035 04:52:14,264 --> 04:52:17,267 WE FOCUSSED IN ON AN LSDM MODEL 6036 04:52:17,267 --> 04:52:21,070 AND I PLUGGED INTO MODELS FOR 6037 04:52:21,070 --> 04:52:23,273 PROPERTIES AND WE USE A COUPLE 6038 04:52:23,273 --> 04:52:25,675 WAYS OF BEING ABLE TO GENERATE 6039 04:52:25,675 --> 04:52:25,942 MOLECULES. 6040 04:52:25,942 --> 04:52:28,545 WE COULD THINK OF THIS APPROACH 6041 04:52:28,545 --> 04:52:31,881 AS BEING AN INTER AND OUTER LOOP 6042 04:52:31,881 --> 04:52:34,551 WHERE IF WE'RE TRYING TO GET 6043 04:52:34,551 --> 04:52:36,719 SELECTIVELY WE'D HAVE THRESHOLDS 6044 04:52:36,719 --> 04:52:39,255 WITH TARGETS AND OFF TARGETS OF 6045 04:52:39,255 --> 04:52:41,257 INTEREST AND GO THROUGH THE 6046 04:52:41,257 --> 04:52:42,926 SELECTIVE WIN TO COME UP WITH 6047 04:52:42,926 --> 04:52:45,061 MOLECULES THAT MASS OUR 6048 04:52:45,061 --> 04:52:45,795 CRITERIA. 6049 04:52:45,795 --> 04:52:47,697 AND ONE EXAMPLE I'LL SHOW IS 6050 04:52:47,697 --> 04:52:50,033 CURRENTLY A MANUSCRIPT THAT'S 6051 04:52:50,033 --> 04:52:50,567 BEEN SUBMITTED. 6052 04:52:50,567 --> 04:52:55,772 THIS IS UNPUBLISHED AS OF NOW 6053 04:52:55,772 --> 04:52:58,541 WHERE WE STARTED USING 6054 04:52:58,541 --> 04:52:59,242 PSILOCYBIN WITH VARIOUS MODELS 6055 04:52:59,242 --> 04:53:01,077 AND RECEPTORS TRYING TO FIND 6056 04:53:01,077 --> 04:53:04,380 MOLECULES THAT WERE AGONISTS AND 6057 04:53:04,380 --> 04:53:06,049 SELECT IT AND ULTIMATELY THE 6058 04:53:06,049 --> 04:53:09,085 MODEL WE GENERATED CAME UP WITH 6059 04:53:09,085 --> 04:53:11,054 01 MOLECULES THAT WE'RE ABLE TO 6060 04:53:11,054 --> 04:53:12,622 SYNTHESIZE AND WE PERFORMED THE 6061 04:53:12,622 --> 04:53:13,590 IN VITRO TESTING AND ULTIMATELY 6062 04:53:13,590 --> 04:53:19,929 ONE OF THESE TURNED OUT TO BE 6063 04:53:19,929 --> 04:53:23,132 SELECTIVE AGONIST WITH AN EC50 6064 04:53:23,132 --> 04:53:30,106 OF ABOUT 12 NANO MOLAR AND FOR 6065 04:53:30,106 --> 04:53:33,776 REFERENCE THERE'S THIS RECEPTOR 6066 04:53:33,776 --> 04:53:38,047 OF 9 NANO MOLMOLAR. 6067 04:53:38,047 --> 04:53:42,585 WE DON'T THINK IT'S THE PERFECT 6068 04:53:42,585 --> 04:53:45,355 MOLECULE BUT HAVE ACHIEVED GOOD 6069 04:53:45,355 --> 04:53:47,657 BRAIN MODELS SO GOING BEYOND 6070 04:53:47,657 --> 04:53:49,659 TRYING TO REPURPOSE THEM BUT 6071 04:53:49,659 --> 04:53:51,361 TRYING TO DESIGN AND SYNTHESIZE 6072 04:53:51,361 --> 04:53:55,064 MOLECULES. 6073 04:53:55,064 --> 04:53:56,966 OBVIOUSLY MORE EXPENSIVE AND 6074 04:53:56,966 --> 04:53:58,368 TIME CONSUMING. 6075 04:53:58,368 --> 04:53:59,769 OUR INTEREST IN THE AREA HAS 6076 04:53:59,769 --> 04:54:02,271 TAKEN OFF IN TERMS OF TRYING TO 6077 04:54:02,271 --> 04:54:04,107 AVOID MOLECULES WITH PSYCHEDELIC 6078 04:54:04,107 --> 04:54:04,340 EFFECTS. 6079 04:54:04,340 --> 04:54:06,542 THIS LED US TO GOING OUT TO 6080 04:54:06,542 --> 04:54:08,277 LITERATURE TRYING TO FIND DATA 6081 04:54:08,277 --> 04:54:09,912 SETS IN VITRO AND IN VIVO WE 6082 04:54:09,912 --> 04:54:12,682 COULD THEN MODEL TO PREDICT 6083 04:54:12,682 --> 04:54:14,550 WHETHER A MOLECULE WILL BE 6084 04:54:14,550 --> 04:54:17,186 PSYCHEDELIC OR NOT. 6085 04:54:17,186 --> 04:54:19,689 A RECENT PUBLICATION HIGHLIGHT 6086 04:54:19,689 --> 04:54:20,056 THE WORK. 6087 04:54:20,056 --> 04:54:23,960 A GREAT EXAMPLE OF GOING INTO 6088 04:54:23,960 --> 04:54:26,929 THE LITERATURE AND TRYING TO 6089 04:54:26,929 --> 04:54:30,566 MANUALLY CREATE DATA FROM BOOKS 6090 04:54:30,566 --> 04:54:34,037 WHERE IT MADE MOLECULES THAT 6091 04:54:34,037 --> 04:54:35,571 WERE PSYCHEDELIC OR NOT AND 6092 04:54:35,571 --> 04:54:37,440 TESTED THEM ON HIMSELF AND 6093 04:54:37,440 --> 04:54:38,541 FRIENDS AND SCORE THEM AND WE 6094 04:54:38,541 --> 04:54:42,578 WERE ABLE TO BUILD 6095 04:54:42,578 --> 04:54:45,081 CLASSIFICATION MODELS AND THEN 6096 04:54:45,081 --> 04:54:46,883 ABLE TO VALIDATE WITH ADDITIONAL 6097 04:54:46,883 --> 04:54:48,051 MODELS AND THESE MODELS IN VITRO 6098 04:54:48,051 --> 04:54:51,087 AND IN VIVO WE USED TO PREDICT 6099 04:54:51,087 --> 04:54:54,090 OTHER MOLECULES AS WELL AS 6100 04:54:54,090 --> 04:54:55,224 MOLECULES WITH OTHER DATA. 6101 04:54:55,224 --> 04:54:56,626 A STARTING POINT TO TRY TO COME 6102 04:54:56,626 --> 04:54:58,861 UP WITH THE NEXT GENERATION OF 6103 04:54:58,861 --> 04:55:02,765 MODELS TO PREDICT PSYCHEDELIC 6104 04:55:02,765 --> 04:55:06,869 EFFECT. 6105 04:55:06,869 --> 04:55:07,770 OBVIOUSLY THERE'S GREAT 6106 04:55:07,770 --> 04:55:09,005 OPPORTUNITY BUT A WORD OF 6107 04:55:09,005 --> 04:55:11,841 CAUTION AND ONE EXAMPLE LED TO A 6108 04:55:11,841 --> 04:55:15,078 PUBLICATION A COUPLE YEARS AGO 6109 04:55:15,078 --> 04:55:16,045 THAT RECEIVED A LITTLE BIT OF 6110 04:55:16,045 --> 04:55:17,747 VISIBILITY AND ENDED UP ON 6111 04:55:17,747 --> 04:55:21,584 NETFLIX WHERE WE WERE WE WERE 6112 04:55:21,584 --> 04:55:23,352 TRYING TO SHOW THE POTENTIAL FOR 6113 04:55:23,352 --> 04:55:24,554 MISUSE OF A.I. 6114 04:55:24,554 --> 04:55:32,095 IN THIS CASE WE'RE DESIGNING NEW 6115 04:55:32,095 --> 04:55:34,564 MOLECULE USING MODELS FOR 6116 04:55:34,564 --> 04:55:35,431 INHIBITION AND OBVIOUSLY THIS IS 6117 04:55:35,431 --> 04:55:36,532 NOT IDEAL. 6118 04:55:36,532 --> 04:55:38,534 WE DIDN'T MAKE THE MOLECULE BUT 6119 04:55:38,534 --> 04:55:42,205 IT GOES TO SHOW THESE KINDS OF 6120 04:55:42,205 --> 04:55:43,106 TECHNOLOGIES CAN DESIGN 6121 04:55:43,106 --> 04:55:44,407 PLAUSIBLE MOLECULES AND THE 6122 04:55:44,407 --> 04:55:47,710 POTENTIAL FOR MISUSE IS GREAT. 6123 04:55:47,710 --> 04:55:50,279 SO JUST TO SUMMARIZE, THERE'S A 6124 04:55:50,279 --> 04:55:52,148 LARGE AMOUNT OF INFORMATION BY 6125 04:55:52,148 --> 04:55:53,649 ACTIVITY FOR TARGETS IN THE 6126 04:55:53,649 --> 04:55:54,951 PUBLIC DOMAIN. 6127 04:55:54,951 --> 04:55:57,453 THIS IS RIFE FOR CURATION AND 6128 04:55:57,453 --> 04:55:57,720 MODELLING. 6129 04:55:57,720 --> 04:55:58,855 HOPEFULLY WHAT I'VE GIVEN IS A 6130 04:55:58,855 --> 04:56:00,890 TASTE OF WHAT WE'VE BEEN ABLE TO 6131 04:56:00,890 --> 04:56:02,558 DO WITH SMALL AMOUNTS OF FUNDING 6132 04:56:02,558 --> 04:56:03,893 IN GENERAL MATERIALS. 6133 04:56:03,893 --> 04:56:08,297 I THINK A.I. ENABLES US TO MOVE 6134 04:56:08,297 --> 04:56:10,199 DEFINITELY FASTER, SAVES US A 6135 04:56:10,199 --> 04:56:11,367 LOT OF MONEY AND WE'RE DOING 6136 04:56:11,367 --> 04:56:14,570 THIS CHEAPLY AND ABLE TO REDUCE 6137 04:56:14,570 --> 04:56:17,707 ANIMAL USE. 6138 04:56:17,707 --> 04:56:21,677 THE CHALLENGE IS WE HAVE TO GO 6139 04:56:21,677 --> 04:56:23,679 FURTHER AND TAKE THE HITS OR 6140 04:56:23,679 --> 04:56:27,083 MOLECULES IN ANIMAL MODELS AND 6141 04:56:27,083 --> 04:56:28,217 IT'S EXPENSIVE. 6142 04:56:28,217 --> 04:56:29,519 DESIGNING IS CHEAP AND FINDING 6143 04:56:29,519 --> 04:56:30,753 MOLECULES IS CHEAP AND TESTING 6144 04:56:30,753 --> 04:56:33,289 AND SYNTHESIZING THEM IS A HUGE 6145 04:56:33,289 --> 04:56:43,432 PROBLEM. 6146 04:56:45,535 --> 04:56:47,503 WE'LL STILL HAVE THIS VALLEY OF 6147 04:56:47,503 --> 04:56:48,204 DEATH. 6148 04:56:48,204 --> 04:56:50,006 A.I. IS ANOTHER TOOL IN OUR TOOL 6149 04:56:50,006 --> 04:56:54,544 KIT AND FOR US IT'S BEEN IMPACT. 6150 04:56:54,544 --> 04:57:01,384 I'D LIKE TO ACKNOWLEDGE THE NIH 6151 04:57:01,384 --> 04:57:03,920 AND THE COLLABORATORS ON EBOLA 6152 04:57:03,920 --> 04:57:05,755 AND TOO MANY COLLABORATORS TO 6153 04:57:05,755 --> 04:57:11,794 ACKNOWLEDGE BUT GRATEFUL FOR THE 6154 04:57:11,794 --> 04:57:12,662 IMPACT AND COLLABORATORS ON 6155 04:57:12,662 --> 04:57:15,932 COVID AND FINALLY MY TEAM WHO 6156 04:57:15,932 --> 04:57:18,234 HAVE BEEN INVOLVED. 6157 04:57:18,234 --> 04:57:26,542 MOST THE WORK I'VE SHOWN 6158 04:57:26,542 --> 04:57:28,511 INVOLVES THESE PEOPLE AND IF 6159 04:57:28,511 --> 04:57:30,379 YOU'D LIKE TO LEARN MORE WE PUT 6160 04:57:30,379 --> 04:57:31,948 TOGETHER A BOOK THAT SHOULD BE 6161 04:57:31,948 --> 04:57:33,082 COMING OUT LATER THIS YEAR. 6162 04:57:33,082 --> 04:57:36,052 I'D LIKE TO THANK OUR OTHER 6163 04:57:36,052 --> 04:57:37,320 COLLABORATORS AND FROM YOU HAVE 6164 04:57:37,320 --> 04:57:41,490 AN A.I. FOCUSSED MANUSCRIPTS, 6165 04:57:41,490 --> 04:57:43,392 PLEASE SUBMIT THEM FOR 6166 04:57:43,392 --> 04:57:45,061 METABOLISM DISPOSITION AND A.I. 6167 04:57:45,061 --> 04:57:45,394 LIFE SCIENCES. 6168 04:57:45,394 --> 04:57:48,397 THANK YOU. 6169 04:57:48,397 --> 04:57:49,332 >> THANK YOU, SEAN FOR AN 6170 04:57:49,332 --> 04:57:55,872 EXCELLENT TALK. 6171 04:57:55,872 --> 04:57:57,940 I'M TRYING TO SEE IF THERE ARE 6172 04:57:57,940 --> 04:57:58,507 ANY QUESTIONS. 6173 04:57:58,507 --> 04:57:59,675 NONE SO FAR. 6174 04:57:59,675 --> 04:58:01,410 MAYBE I CAN ASK ONE QUESTION, 6175 04:58:01,410 --> 04:58:04,881 SEAN. 6176 04:58:04,881 --> 04:58:09,952 THE STUDIES SHOWED PARTICULARLY 6177 04:58:09,952 --> 04:58:18,561 FOR EBOLA, I ASSUME IS THAT 6178 04:58:18,561 --> 04:58:22,064 MANUFACTURED AT THIS POINT OR 6179 04:58:22,064 --> 04:58:24,267 MORE STILL IN THE LAB 6180 04:58:24,267 --> 04:58:25,568 DEVELOPMENT STAGES? 6181 04:58:25,568 --> 04:58:27,036 >> IT'S ACTUALLY AN APPROVED 6182 04:58:27,036 --> 04:58:31,374 DRUG IN EUROPE. 6183 04:58:31,374 --> 04:58:33,075 IT'S USED IN AFRICA. 6184 04:58:33,075 --> 04:58:36,312 IT'S USED AS AN ANTI-MARLARIAL 6185 04:58:36,312 --> 04:58:38,547 BUT NOT AN INDIVIDUAL MOLECULE. 6186 04:58:38,547 --> 04:58:45,087 WE WERE SHOWING THE INDIVIDUAL 6187 04:58:45,087 --> 04:58:55,564 USE AND YOU CAN SEE ONE THE 6188 04:59:17,219 --> 04:59:17,520 GROUP. 6189 04:59:17,520 --> 04:59:19,422 I'M HAPPY TO TOUCH BASE WITH YOU 6190 04:59:19,422 --> 04:59:20,323 LATER AND FOLLOW-UP. 6191 04:59:20,323 --> 04:59:21,424 >> THAT WOULD BE GREAT. 6192 04:59:21,424 --> 04:59:22,558 THANK YOU. 6193 04:59:22,558 --> 04:59:24,293 WE PRESENTED IT TO BARDA OVER 6194 04:59:24,293 --> 04:59:24,994 THE YEARS. 6195 04:59:24,994 --> 04:59:29,565 I THINK WHEN THERE WAS THE COVID 6196 04:59:29,565 --> 04:59:30,299 O 6197 04:59:30,299 --> 04:59:31,701 OUTBREAK SEVERAL TIMES AND GOT 6198 04:59:31,701 --> 04:59:33,069 NO INTEREST FROM ANY OF THE 6199 04:59:33,069 --> 04:59:34,537 ORGANIZATIONS WE PRESENTED TO. 6200 04:59:34,537 --> 04:59:36,172 I THINK THEY WERE LOOKING FOR 6201 04:59:36,172 --> 04:59:39,575 NOVEL MOLECULES AND LESS ON THE 6202 04:59:39,575 --> 04:59:40,843 REPURPOSING SIDE. 6203 04:59:40,843 --> 04:59:43,079 THERE'S DEFINITELY PROS AND CONS 6204 04:59:43,079 --> 04:59:47,049 TO REPURPOSING AND WE WERE 6205 04:59:47,049 --> 04:59:49,018 CERTAINLY INTERESTED TO SEE WHAT 6206 04:59:49,018 --> 04:59:50,353 INITIATIVES WERE OUT THERE AND 6207 04:59:50,353 --> 04:59:52,121 AS A SMALL COMPANY SOME REQUIRED 6208 04:59:52,121 --> 04:59:55,257 US TO PUT UP FUNDING AND 6209 04:59:55,257 --> 04:59:57,426 CERTAINLY DIDN'T HAVE THE WHE 6210 04:59:57,426 --> 05:00:02,098 WHEREWITHAL AND THAT WAS A 6211 05:00:02,098 --> 05:00:03,866 CHALLENGE IS A SIGNIFICANT 6212 05:00:03,866 --> 05:00:05,134 INVESTMENT FROM OUR SIDE AND 6213 05:00:05,134 --> 05:00:08,037 THAT'S IMPOSSIBLE FOR A TINY 6214 05:00:08,037 --> 05:00:11,040 COMPANY SO IT WASN'T REALISTIC. 6215 05:00:11,040 --> 05:00:11,807 THANK YOU. 6216 05:00:11,807 --> 05:00:13,976 >> BASED ON WHAT YOU PRESENTED, 6217 05:00:13,976 --> 05:00:18,547 ALL THE OTHER MOLECULES IN THE 6218 05:00:18,547 --> 05:00:26,489 GPCR MODULATORS IS THAT FOCUS TO 6219 05:00:26,489 --> 05:00:28,891 LOOK AT ALREADY ESTABLISHED 6220 05:00:28,891 --> 05:00:31,627 MOLECULES AND REPOSITION THEM? 6221 05:00:31,627 --> 05:00:34,563 IT SEEMS LIKE YOUR A.I. TOOLS 6222 05:00:34,563 --> 05:00:36,699 ARE PREDOMINANTLY DIRECTED 6223 05:00:36,699 --> 05:00:37,466 TOWARDS THAT RATHER THAN 6224 05:00:37,466 --> 05:00:38,000 DEVELOPING NEW MOLECULES. 6225 05:00:38,000 --> 05:00:39,568 IS THAT A FAIR STATEMENT? 6226 05:00:39,568 --> 05:00:42,705 >> WHAT I WAS TRYING TO DO AND 6227 05:00:42,705 --> 05:00:43,672 PROBABLY WASN'T CLEAR ABOUT IT 6228 05:00:43,672 --> 05:00:45,808 WHICH I'M HAPPY TO CLARIFY SO 6229 05:00:45,808 --> 05:00:47,977 IT'S A GREAT QUESTION. 6230 05:00:47,977 --> 05:00:53,949 WE STARTED OFF ON REPURPOSING 6231 05:00:53,949 --> 05:00:56,685 HEAVILY ON THE TRADITIONAL A.I. 6232 05:00:56,685 --> 05:00:58,421 JUST TRYING TO MINE DATA SETS 6233 05:00:58,421 --> 05:01:01,557 AND MOVED TO DESIGNING NEW 6234 05:01:01,557 --> 05:01:02,158 MOLECU 6235 05:01:02,158 --> 05:01:02,425 MOLECULES. 6236 05:01:02,425 --> 05:01:03,092 MORE DE NOVO DESIGN AND I THINK 6237 05:01:03,092 --> 05:01:05,094 THAT'S WHERE WE ARE TODAY. 6238 05:01:05,094 --> 05:01:09,598 WE'RE MORE ON THAT DE NOVO 6239 05:01:09,598 --> 05:01:17,239 DESIGN SIDE. 6240 05:01:17,239 --> 05:01:19,475 WE STILL SCREEN THROUGH BUT WE 6241 05:01:19,475 --> 05:01:21,243 HAD SO MUCH FEEDBACK FROM PHARMA 6242 05:01:21,243 --> 05:01:22,945 AND THERE WERE ONLY INTERESTED 6243 05:01:22,945 --> 05:01:26,348 IN MOFL MOLECULES. 6244 05:01:26,348 --> 05:01:32,321 THEY WEREN'T INTERESTED IN 6245 05:01:32,321 --> 05:01:34,557 REPURPOSING AND THAT'S WHAT 6246 05:01:34,557 --> 05:01:37,693 DROVE US TO DO DESIGN AND FIND 6247 05:01:37,693 --> 05:01:39,562 FUNDERS TO DO THE SYNTHESIS. 6248 05:01:39,562 --> 05:01:46,535 WE'VE BEEN FORTUNATE TO GET 6249 05:01:46,535 --> 05:01:48,471 GRANTS AND NIDA THAT ALLOWED TO 6250 05:01:48,471 --> 05:01:50,906 DO MORE GENERATIVE DESIGN AND 6251 05:01:50,906 --> 05:01:52,208 COMING UP WITH DESIGNS FOR 6252 05:01:52,208 --> 05:01:55,444 OPIOID USE DISORDER AND HAVE 6253 05:01:55,444 --> 05:01:57,079 BEEN AUTOMOBILE TO LEVERAGE 6254 05:01:57,079 --> 05:01:59,815 INHIBITION WORK TO TRY TO DESIGN 6255 05:01:59,815 --> 05:02:01,817 MOLECULES AS POTENTIAL 6256 05:02:01,817 --> 05:02:04,386 REACTIVATORS USED FOR PESTICIDE 6257 05:02:04,386 --> 05:02:06,555 AND CHEMICAL WEAPON POISONING. 6258 05:02:06,555 --> 05:02:10,559 IT'S EVOLVED AND THERE'S BEEN 6259 05:02:10,559 --> 05:02:16,532 SERIOUS PAIR PIVOTS BUT HAPPY 6260 05:02:16,532 --> 05:02:17,833 TO FOLLOW-UP WITH ANYONE IF YOU 6261 05:02:17,833 --> 05:02:26,008 WANT TO DROP ME AN E-MAIL. 6262 05:02:26,008 --> 05:02:29,478 >> INSTEAD OF SYNTHESIZING 6263 05:02:29,478 --> 05:02:32,348 COMPOUNDS, DO YOU HAVE GOOD LUCK 6264 05:02:32,348 --> 05:02:34,550 PURCHASING COMPOUNDS THAT ARE IN 6265 05:02:34,550 --> 05:02:38,888 SIMILARLY TO THE VIRTUAL HITS? 6266 05:02:38,888 --> 05:02:40,656 >> WE'VE TRIED THAT AS WELL. 6267 05:02:40,656 --> 05:02:43,092 I THINK THE MOST CONVINCING -- 6268 05:02:43,092 --> 05:02:45,227 THIS IS JUST REALLY A FIS SOF 6269 05:02:45,227 --> 05:02:53,903 CAL -- PHILOSOPHICAL QUESTION. 6270 05:02:53,903 --> 05:02:54,570 I'VE BEEN INTERESTED TO VALIDATE 6271 05:02:54,570 --> 05:02:56,572 THE MODELS CONVINCINGLY AND YOU 6272 05:02:56,572 --> 05:02:59,074 CAN NEVER REALLY WIN OVER PEOPLE 6273 05:02:59,074 --> 05:03:00,409 UNLESS YOU DESIGN NOVEL 6274 05:03:00,409 --> 05:03:00,676 MOLECULES. 6275 05:03:00,676 --> 05:03:02,378 BUYING THEM IS A CHEAP WAY TO DO 6276 05:03:02,378 --> 05:03:03,746 IT AND I'M FINE WITH THAT BUT 6277 05:03:03,746 --> 05:03:06,448 PEOPLE ARE NOT GOING TO BE 6278 05:03:06,448 --> 05:03:09,885 REALLY TRULY CONVINCED UNTIL YOU 6279 05:03:09,885 --> 05:03:11,654 DESIGN TRULY NOVEL MOLECULE THAT 6280 05:03:11,654 --> 05:03:13,355 ARE NOT AVAILABLE FROM ANYONE 6281 05:03:13,355 --> 05:03:14,523 ELSE, HAVE NEVER BEEN PUBLISHED, 6282 05:03:14,523 --> 05:03:16,392 NOT PUT INTO A PATENT. 6283 05:03:16,392 --> 05:03:17,393 THAT'S THE ONLY WAY WE'LL 6284 05:03:17,393 --> 05:03:19,328 CONVINCE PEOPLE. 6285 05:03:19,328 --> 05:03:22,531 AND THEN YOU JUST GET IN THE MIN 6286 05:03:22,531 --> 05:03:30,105 YOU -- MINUTIA AND IT'S NEVER 6287 05:03:30,105 --> 05:03:31,540 ENDING AND SOMEONE WILL FIND A 6288 05:03:31,540 --> 05:03:36,845 METHOD OR SIMILARITY OR WHATEVER 6289 05:03:36,845 --> 05:03:41,050 OH, IT'S SIMILAR TO THIS AND 6290 05:03:41,050 --> 05:03:42,851 THERE'S -- IT'S CHEAPER TO DO 6291 05:03:42,851 --> 05:03:44,653 THAT BUT THE MOST CONVINCING 6292 05:03:44,653 --> 05:03:46,188 THING IS TO MAKE MOLECULES. 6293 05:03:46,188 --> 05:03:48,557 >> IF YOU CAN STAY ONLINE, SEAN 6294 05:03:48,557 --> 05:03:51,660 AND ANSWER THE REST OF THE 6295 05:03:51,660 --> 05:03:52,895 QUESTION. 6296 05:03:52,895 --> 05:03:57,066 >> THANK YOU. 6297 05:03:57,066 --> 05:04:01,537 >> OUR NEXT SPEAKER IS WOODY 6298 05:04:01,537 --> 05:04:01,971 SHERMAN. 6299 05:04:01,971 --> 05:04:05,774 HE IS A CHIEF INNOVATION OFFICER 6300 05:04:05,774 --> 05:04:09,211 OF A THERAPEUTICS COMPANY AND 6301 05:04:09,211 --> 05:04:12,881 WAS CHIEF COMPUTATIONAL 6302 05:04:12,881 --> 05:04:14,283 SCIENTIST AND AT SILICON 6303 05:04:14,283 --> 05:04:17,486 THERAPEUTICS AND HEAD OF 6304 05:04:17,486 --> 05:04:19,622 APPLICATIONS SCIENCE AT 6305 05:04:19,622 --> 05:04:19,922 SHORTINGER. 6306 05:04:19,922 --> 05:04:24,193 AND HE'S A PROFESSOR AT THE 6307 05:04:24,193 --> 05:04:25,427 UNIVERSITY OF MASSACHUSETTS 6308 05:04:25,427 --> 05:04:29,632 AMHERST WHERE HE GIVES LECTURES 6309 05:04:29,632 --> 05:04:31,467 ON DRUG DESIGN AND COMPANY 6310 05:04:31,467 --> 05:04:31,734 CREATION. 6311 05:04:31,734 --> 05:04:35,137 HE COMPLETED HIS Ph.D. AT MIT 6312 05:04:35,137 --> 05:04:41,410 AND PUBLISHED 400 PEER REVIEWED 6313 05:04:41,410 --> 05:04:43,746 PAPERS INCLUDING QUANTUM 6314 05:04:43,746 --> 05:04:45,648 MECHANICS AND OTHER AREAS OF 6315 05:04:45,648 --> 05:04:46,682 INFORMATICS AND MACHINE 6316 05:04:46,682 --> 05:04:47,816 LEARNING. 6317 05:04:47,816 --> 05:04:50,552 HIS TOPIC IS FROM CONCEPT TO 6318 05:04:50,552 --> 05:04:58,260 CLINIC. 6319 05:04:58,260 --> 05:04:58,560 DISCOVERY. 6320 05:04:58,560 --> 05:05:01,030 THE FLOOR IS ALL YOURS. 6321 05:05:01,030 --> 05:05:02,831 >> THANK YOU FOR THE 6322 05:05:02,831 --> 05:05:03,465 INTRODUCTION AND THANK YOU FOR 6323 05:05:03,465 --> 05:05:08,671 THE INVITATION TO BE HERE AND 6324 05:05:08,671 --> 05:05:08,971 ORGANIZERS. 6325 05:05:08,971 --> 05:05:10,539 ARE YOU SEEING THE FULL SCREEN 6326 05:05:10,539 --> 05:05:13,709 MODE AND NOT MY PRESENTATION 6327 05:05:13,709 --> 05:05:14,209 MODE? 6328 05:05:14,209 --> 05:05:14,543 >> YES. 6329 05:05:14,543 --> 05:05:15,077 >> GREAT. 6330 05:05:15,077 --> 05:05:21,116 I'LL START WITH A HIGH LEVEL 6331 05:05:21,116 --> 05:05:26,488 SUMMARY OF OUR VIEW AND IT'S AN 6332 05:05:26,488 --> 05:05:27,890 A.I. SYMPOSIUM AND JOHN DID A 6333 05:05:27,890 --> 05:05:29,124 GREAT JOB EMPHASIZING THE 6334 05:05:29,124 --> 05:05:32,461 IMPORTANCE OF DYNAMICS SO I'LL 6335 05:05:32,461 --> 05:05:37,232 BE ABLE TO GET THROUGH THIS 6336 05:05:37,232 --> 05:05:37,566 QUICKLY. 6337 05:05:37,566 --> 05:05:39,101 TALK ABOUT THE PLATFORM WE'VE 6338 05:05:39,101 --> 05:05:43,806 BEEN DEVELOPING SINCE THE DAYS 6339 05:05:43,806 --> 05:05:46,909 OF SILICON THERAPEUTICS AND 6340 05:05:46,909 --> 05:05:57,386 CONTINUE TO DEVELOP WITHIN 6341 05:06:03,592 --> 05:06:05,027 PSIVANT AND LIVING THINGS CAN BE 6342 05:06:05,027 --> 05:06:06,895 UNDERSTAND IN TERMS OF THE 6343 05:06:06,895 --> 05:06:08,764 JIGGLING OF ATOMS. 6344 05:06:08,764 --> 05:06:10,199 WE'VE KNOWN THE CHALLENGE IS 6345 05:06:10,199 --> 05:06:11,433 BEING ABLE TO MAKE PREDICTIONS 6346 05:06:11,433 --> 05:06:14,603 AN RUN SIMULATIONS AT SCALE TO 6347 05:06:14,603 --> 05:06:17,005 INFLUENCE OUR DRUG DISCOVERY 6348 05:06:17,005 --> 05:06:18,540 PROCESSES. 6349 05:06:18,540 --> 05:06:29,017 WE STARTED WITH THE LAB AND 6350 05:06:35,691 --> 05:06:37,192 SOMETIMES COMPUTRATION IS THE 6351 05:06:37,192 --> 05:06:39,094 RIGHT TOOL AND BEING PRACTICAL 6352 05:06:39,094 --> 05:06:42,564 AND FOCUSSING ON WHAT MATTERS IN 6353 05:06:42,564 --> 05:06:46,068 CONTEXT OF OUR DRUG DISCOVERY 6354 05:06:46,068 --> 05:06:46,435 PRODUCT. 6355 05:06:46,435 --> 05:06:49,171 WE'RE NOT JUST TALKING ABOUT 6356 05:06:49,171 --> 05:06:51,540 SIMULATIONS AND WATCHING THINGS 6357 05:06:51,540 --> 05:06:53,409 JIGGLE AROUND WE'RE INTERESTED 6358 05:06:53,409 --> 05:06:54,777 IN FREE ENERGY LANDSCAPES AND 6359 05:06:54,777 --> 05:06:58,347 DIFFERENT CONFIRMATIONS AND HOW 6360 05:06:58,347 --> 05:07:00,482 THOSE RELATE TO BIOLOGY. 6361 05:07:00,482 --> 05:07:07,055 WE DON'T DO THIS WITH MOLECULAR 6362 05:07:07,055 --> 05:07:08,257 DYNAMICS AND BELIEVE THAT IS 6363 05:07:08,257 --> 05:07:09,792 GOING TO BE THE MOST EFFECTIVE 6364 05:07:09,792 --> 05:07:14,763 WAY OF MAKING AN IMPACT WITH 6365 05:07:14,763 --> 05:07:16,165 MOLECULAR SIMULATIONS. 6366 05:07:16,165 --> 05:07:17,499 HOW DO WE THINK ABOUT THIS? 6367 05:07:17,499 --> 05:07:20,302 WITH KNOW PROTEINS ARE FLEXIBLE 6368 05:07:20,302 --> 05:07:22,538 AND THAT FLEXIBILITY ENDS UP 6369 05:07:22,538 --> 05:07:23,939 INFLUENCING BIOLOGY. 6370 05:07:23,939 --> 05:07:25,974 HOW WE ARE HANDLING THESE WITH 6371 05:07:25,974 --> 05:07:27,443 THE SIMULATIONS WE DEAL WITH 6372 05:07:27,443 --> 05:07:30,212 WATERS AND COFACTORS AND IONS 6373 05:07:30,212 --> 05:07:32,514 AND THE THINGS IN THERE. 6374 05:07:32,514 --> 05:07:33,715 OUR ULTIMATE GOAL IS 6375 05:07:33,715 --> 05:07:35,017 UNDERSTANDING HOW MOLECULES 6376 05:07:35,017 --> 05:07:36,084 BOUND TO PROTEINS. 6377 05:07:36,084 --> 05:07:37,553 NOT JUST HOW THEY BIND IN TERMS 6378 05:07:37,553 --> 05:07:39,655 OF BINDING AFFINITIES BUT WHAT 6379 05:07:39,655 --> 05:07:42,057 HAPPENS WHEN THEY BIND TO THE 6380 05:07:42,057 --> 05:07:46,528 PROTEIN CONFIRMATIONAL CHANGES 6381 05:07:46,528 --> 05:07:48,897 AS MOLECULES LOCK INTO CERTAIN 6382 05:07:48,897 --> 05:07:50,532 CONFIRMATIONS AND HOW TO GO 6383 05:07:50,532 --> 05:07:53,502 ABOUT DESIGNING BETTER MOLECULE 6384 05:07:53,502 --> 05:07:54,536 WITH BETTER PROFILES AND 6385 05:07:54,536 --> 05:07:55,270 INFLUENCE PROTEIN IN WAYS THAT 6386 05:07:55,270 --> 05:08:00,676 INFLUENCE BIOLOGY. 6387 05:08:00,676 --> 05:08:02,411 WE USE THE SIMULATION TO 6388 05:08:02,411 --> 05:08:04,480 UNDERSTAND MOLECULAR MECHANISMS 6389 05:08:04,480 --> 05:08:10,486 WHICH ARE BASED ON DYNAMICS TO 6390 05:08:10,486 --> 05:08:11,553 PREDICT BINDING POCKETS AND 6391 05:08:11,553 --> 05:08:13,121 STICKS TO A PROTEIN. 6392 05:08:13,121 --> 05:08:16,525 YOU HAVE TO UNDERSTAND NOT JUST 6393 05:08:16,525 --> 05:08:17,192 BINDABILITY BUT DRUGABILITY AND 6394 05:08:17,192 --> 05:08:20,996 LOU THEY RELATE TO FUNCTION 6395 05:08:20,996 --> 05:08:24,666 BASED ON DYNAMICS AND BINDING 6396 05:08:24,666 --> 05:08:26,001 FREE ENERGY CALCULATION WHICH I 6397 05:08:26,001 --> 05:08:27,302 BELIEVE ARE THE MOST ACCURATE 6398 05:08:27,302 --> 05:08:35,077 APPROACH TO BINDING AFFINITY 6399 05:08:35,077 --> 05:08:38,580 ALSO DONE WITH MOLECULAR DYN 6400 05:08:38,580 --> 05:08:42,384 DYNAMICS EXAMPLE. 6401 05:08:42,384 --> 05:08:43,852 WHY NOT MACHINE LEARNING FOR 6402 05:08:43,852 --> 05:08:44,152 EVERYTHING? 6403 05:08:44,152 --> 05:08:47,389 WE'RE A DATA POOR REGIME. 6404 05:08:47,689 --> 05:08:50,559 WE NEED TO WORK ON THAT IT'S A 6405 05:08:50,559 --> 05:09:00,102 GREAT PLACE TO PITCH IN THERE'S 6406 05:09:00,102 --> 05:09:02,504 BEEN PROTEIN STRUCTURE 6407 05:09:02,504 --> 05:09:10,345 PREDICTION AND THAT'S ONE OF 6408 05:09:10,345 --> 05:09:11,280 MANY AREAS TO SOLVE. 6409 05:09:11,280 --> 05:09:12,781 YOU WE'RE TOLD WHAT TO DO BY 6410 05:09:12,781 --> 05:09:14,550 PEOPLE AND YOU NEED TO PROVIDE 6411 05:09:14,550 --> 05:09:18,053 INSIGHT TO PEOPLE AND THERE'S 6412 05:09:18,053 --> 05:09:19,087 NOT A LOT OF IN THAT SPACE 6413 05:09:19,087 --> 05:09:20,255 THOUGH IT'S A GREAT AREA OF 6414 05:09:20,255 --> 05:09:21,957 RESEARCH FOLKS ARE WORKING ON 6415 05:09:21,957 --> 05:09:23,859 AND THERE'S NOT A LOT OF END TO 6416 05:09:23,859 --> 05:09:25,227 END SOLUTIONS PVEN I THINK 6417 05:09:25,227 --> 05:09:27,095 SEVERAL FOLKS HERE ARE TALKED 6418 05:09:27,095 --> 05:09:31,433 ABOUT BUILDING ALONG THAT 6419 05:09:31,433 --> 05:09:31,667 PROCESS. 6420 05:09:31,667 --> 05:09:33,101 WE'RE STILL EARLY IN THE DAYS 6421 05:09:33,101 --> 05:09:34,903 NOT TO SAY THEY'RE NOT USEFUL. 6422 05:09:34,903 --> 05:09:37,906 ALPHA FOLD MADE A BIG IMPACT AND 6423 05:09:37,906 --> 05:09:39,374 ESSENTIALLY SOLVED THE PROTEIN 6424 05:09:39,374 --> 05:09:41,376 STRUCTURE PREDICTION PROBLEM BUT 6425 05:09:41,376 --> 05:09:42,578 WITHOUT COMPOUNDS. 6426 05:09:42,578 --> 05:09:44,580 AS SUCH THIS IS CURRENTLY 6427 05:09:44,580 --> 05:09:50,552 COMPARABLE TO SOLVING AN APOE 6428 05:09:50,552 --> 05:09:54,556 CRYSTAL STRUCTURE AND IT MAKES 6429 05:09:54,556 --> 05:09:56,792 MINIMAL IMPACT AND NEED TO TALK 6430 05:09:56,792 --> 05:10:00,696 ABOUT HOW MOLECULES BIND THE 6431 05:10:00,696 --> 05:10:06,635 PROTEINS AND CAN INFLUENCE THE 6432 05:10:06,635 --> 05:10:07,336 STRUCTURE. 6433 05:10:07,336 --> 05:10:09,338 THERE'S GENERATIVE MODELLING AND 6434 05:10:09,338 --> 05:10:10,539 TOOLS OUT THERE QUITE USEFUL IN 6435 05:10:10,539 --> 05:10:13,575 THE CONTEXT OF A BROADER SUITE 6436 05:10:13,575 --> 05:10:14,543 OF THINGS. 6437 05:10:14,543 --> 05:10:17,913 ACTIVE LEARNING. 6438 05:10:17,913 --> 05:10:19,781 CONNOR DEVELOPED THIS TOOL AND 6439 05:10:19,781 --> 05:10:22,084 WE'RE BIG BELIEVERS IN ACTIVE 6440 05:10:22,084 --> 05:10:22,351 LEARNING. 6441 05:10:22,351 --> 05:10:24,219 IT IMPROVES THE EFFICIENCY OF 6442 05:10:24,219 --> 05:10:25,053 WHATEVER YOU'RE APPLYING THE 6443 05:10:25,053 --> 05:10:26,521 SCORING FUNCTION IN. 6444 05:10:26,521 --> 05:10:27,756 WHETHER IT'S DOCKING OR FREE 6445 05:10:27,756 --> 05:10:30,192 ENERGY CALCULATIONS. 6446 05:10:30,192 --> 05:10:32,427 IT DIDN'T MAKE THE CALCULATIONS 6447 05:10:32,427 --> 05:10:34,096 MORE ACCURATE IT JUST ALLOWS YOU 6448 05:10:34,096 --> 05:10:36,264 TO GET TO MORE MOLECULES MORE 6449 05:10:36,264 --> 05:10:40,535 QUICKLY BUT IT'S VALUABLE. 6450 05:10:40,535 --> 05:10:41,203 OVER ALL THERE ARE GREAT TOOLS 6451 05:10:41,203 --> 05:10:43,071 OUT THERE AND THE FIELD IS 6452 05:10:43,071 --> 05:10:48,410 EVOLVING AND WE'RE IN A FASTLY 6453 05:10:48,410 --> 05:10:50,579 GROWING AID WITH INVESTMENTS BUT 6454 05:10:50,579 --> 05:10:53,615 LIMITED PREDICTIVE ABILITY FOR 6455 05:10:53,615 --> 05:10:54,182 LIGAND BINDING. 6456 05:10:54,182 --> 05:10:56,451 ONE QUESTION EARLIER AND THE 6457 05:10:56,451 --> 05:10:59,354 CHAT WAS RELATED TO ML METHODS 6458 05:10:59,354 --> 05:11:01,690 AND THEY CAN PREDICT LIGAND 6459 05:11:01,690 --> 05:11:03,125 AFFINITIES WHEN YOU HAVE A LOT 6460 05:11:03,125 --> 05:11:04,426 OF DATA BUT THEN IT'S LESS 6461 05:11:04,426 --> 05:11:06,595 INTERESTING TO PREDICT AFFINITY 6462 05:11:06,595 --> 05:11:10,832 SO WHEN YOU GET IN THE MOST 6463 05:11:10,832 --> 05:11:12,334 INTERESTING NOVEL DRUG DISCOVERY 6464 05:11:12,334 --> 05:11:14,403 PROBLEMS FOR CHALLENGING TARGETS 6465 05:11:14,403 --> 05:11:16,038 WHERE THERE MAY NOT BE A LOT OF 6466 05:11:16,038 --> 05:11:17,339 CHEMICAL MATTER, PHYSICS IS 6467 05:11:17,339 --> 05:11:21,209 STILL WHAT WE RELY ON MOST 6468 05:11:21,209 --> 05:11:21,910 HEAVILY. 6469 05:11:21,910 --> 05:11:23,845 WE USE ML MODELS AND FOR 6470 05:11:23,845 --> 05:11:24,546 PREDICTIONS AND SOME GENERATIVE 6471 05:11:24,546 --> 05:11:28,483 MODELLING AS WELL. 6472 05:11:28,483 --> 05:11:33,055 OUR APPROACH HAS BEEN SINCE THE 6473 05:11:33,055 --> 05:11:34,156 BEGINNING SINCE THE SILICON 6474 05:11:34,156 --> 05:11:35,957 THERAPEUTICS DAYS TO BUILD A 6475 05:11:35,957 --> 05:11:36,591 TOOL KIT. 6476 05:11:36,591 --> 05:11:38,393 ONE THING I LEARNED MANY TIMES 6477 05:11:38,393 --> 05:11:40,629 OVER THE YEARS IS DRUG DISCOVERY 6478 05:11:40,629 --> 05:11:42,631 IS HARD FOR MANY DIFFERENT 6479 05:11:42,631 --> 05:11:44,466 REASONS AND EACH DRUG DISCOVERY 6480 05:11:44,466 --> 05:11:46,568 TENDS FOR CHALLENGING FOR 6481 05:11:46,568 --> 05:11:47,169 DIFFERENT REASONS. 6482 05:11:47,169 --> 05:11:50,739 WHILE THERE'S SOME RECURRING 6483 05:11:50,739 --> 05:11:53,642 THEMES LIKE BINDING ENERGIES AND 6484 05:11:53,642 --> 05:11:55,310 THAT WAS A UNIQUE PROBLEM THOUGH 6485 05:11:55,310 --> 05:11:58,046 WE WERE TRYING TO PREDICT 6486 05:11:58,046 --> 05:11:59,614 BINDING ENERGIES AND STANDARD 6487 05:11:59,614 --> 05:11:59,815 STUFF. 6488 05:11:59,815 --> 05:12:04,152 RATHER THAN BUILDING A BLACK BOX 6489 05:12:04,152 --> 05:12:06,021 OR SINGLE TOOL TO SOLVE ONE 6490 05:12:06,021 --> 05:12:10,225 THING REALLY WELL, WE BUILT A 6491 05:12:10,225 --> 05:12:10,826 FLEXIBLE TOOLKIT BASED ON 6492 05:12:10,826 --> 05:12:15,330 QUANTUM MECHANICS AND MEC LAR 6493 05:12:15,330 --> 05:12:16,498 SIMULATION AND A.I. MACHINE 6494 05:12:16,498 --> 05:12:19,334 LEARNING AND ROBUST COMPUTING 6495 05:12:19,334 --> 05:12:20,402 INFRASTRUCTURE AND WE HAVE 6496 05:12:20,402 --> 05:12:21,236 PUBLISHED A LOT OF THIS WORK 6497 05:12:21,236 --> 05:12:25,107 ALONG THE WAY. 6498 05:12:25,107 --> 05:12:26,875 FORTUNATELY I DON'T NEED TO TALK 6499 05:12:26,875 --> 05:12:29,845 ABOUT BINDING FREE ENERGY 6500 05:12:29,845 --> 05:12:30,579 METHODS. 6501 05:12:30,579 --> 05:12:33,715 THEY DESCRIBED THE IMPORTANCE OF 6502 05:12:33,715 --> 05:12:34,983 THESE IN THE FREE ENERGY 6503 05:12:34,983 --> 05:12:35,951 CONSORTIUM IS EXCELLENT AND 6504 05:12:35,951 --> 05:12:38,520 WE'RE STRONG SUPPORTERS OF THE 6505 05:12:38,520 --> 05:12:39,921 OPEN SCIENCE MOVEMENT IN GENERAL 6506 05:12:39,921 --> 05:12:41,757 AS I'LL TALK ABOUT BUT FOR FREE 6507 05:12:41,757 --> 05:12:44,559 ENERGY CALCULATION WE BUILT OUR 6508 05:12:44,559 --> 05:12:45,694 OWN INTERNAL INFRASTRUCTURE AND 6509 05:12:45,694 --> 05:12:48,263 BUILD ON TOOLS ARE OUT THERE 6510 05:12:48,263 --> 05:12:50,599 WHEN THEY ARE EXIST FOR EXAMPLE 6511 05:12:50,599 --> 05:12:52,334 STARTING WITH FLOW MAP TOOL FOR 6512 05:12:52,334 --> 05:12:54,069 CERTAIN PIECES OF IT. 6513 05:12:54,069 --> 05:12:55,270 WE HAD TO BUILD A LOT OF STUFF 6514 05:12:55,270 --> 05:12:57,606 ON OUR OWN AND INTEGRATE THIS. 6515 05:12:57,606 --> 05:12:59,841 THE REASON WE'VE DONE THIS IS WE 6516 05:12:59,841 --> 05:13:01,843 CAN RUN AT SCALE WHICH IS SUPER 6517 05:13:01,843 --> 05:13:03,145 IMPORTANT FOR ASKING BIG 6518 05:13:03,145 --> 05:13:04,212 QUESTIONS ABOUT NOT JUST LOOKING 6519 05:13:04,212 --> 05:13:06,481 AT A HANDFUL OF MOLECULE BUT 6520 05:13:06,481 --> 05:13:08,583 HUNDREDS OR THOUSANDS ON A 6521 05:13:08,583 --> 05:13:09,117 ROUTINE BASIS. 6522 05:13:09,117 --> 05:13:12,721 AND ALLOWS US TO CUSTOMIZE THE 6523 05:13:12,721 --> 05:13:14,556 TOOLS TO SOLVE OUR PROBLEMS AND 6524 05:13:14,556 --> 05:13:17,292 EXPLAIN WHY THAT WAS IMPORTANT. 6525 05:13:17,292 --> 05:13:18,560 WE BUILT OUT OTHER FUNCTIONALITY 6526 05:13:18,560 --> 05:13:21,763 THAT IS ABOUT ENABLING HUMANS. 6527 05:13:21,763 --> 05:13:23,698 WE DON'T JUST WANT TO PRODUCE A 6528 05:13:23,698 --> 05:13:26,301 NUMBER BUT ADDITIONAL INSIGHT AS 6529 05:13:26,301 --> 05:13:26,568 WELL. 6530 05:13:26,568 --> 05:13:29,104 WE'VE CREATED THIS FREE ENERGY 6531 05:13:29,104 --> 05:13:31,072 COMPOSITION WHERE YOU CAN 6532 05:13:31,072 --> 05:13:34,176 EXTRACT FROM THE FREE ENERGY 6533 05:13:34,176 --> 05:13:35,076 SIMULATIONS ATOMIC 6534 05:13:35,076 --> 05:13:35,410 CONTRIBUTIONS. 6535 05:13:35,410 --> 05:13:37,813 THIS IS NOT RIGOROUS BUT USEFUL. 6536 05:13:37,813 --> 05:13:40,782 YOU CAN'T BREAK UP FREE ENERGIES 6537 05:13:40,782 --> 05:13:45,287 IN TO ATOMIC CONTRIBUTIONS WITH 6538 05:13:45,287 --> 05:13:47,022 ENTROPY IN THERE AND THE 6539 05:13:47,022 --> 05:13:52,294 MOLECULE COULD BE GOOD BY 6540 05:13:52,294 --> 05:13:53,628 PREDICTIVE MINUS AND THE THINGS 6541 05:13:53,628 --> 05:13:55,030 YOU MIGHT NOT WANT TO MESS WITH 6542 05:13:55,030 --> 05:13:57,966 AND OTHER PARTS OF THE LIGAND 6543 05:13:57,966 --> 05:13:59,267 THAT CHSHOULD BE CHANGED AND WH 6544 05:13:59,267 --> 05:14:00,969 IS UNHAPPY TO LEAD TO ADDITIONAL 6545 05:14:00,969 --> 05:14:02,771 DESIGN INSIGHT. 6546 05:14:02,771 --> 05:14:05,607 AND IN ADDITION TO BINDING FREE 6547 05:14:05,607 --> 05:14:06,308 ENERGY CALCULATION WE DEVELOPED 6548 05:14:06,308 --> 05:14:08,677 A LOT IN THE CONTEXT OF WHAT WE 6549 05:14:08,677 --> 05:14:13,014 CALL BIO LOGICALLY RELEVANT 6550 05:14:13,014 --> 05:14:14,583 PROTEIN MOTION. 6551 05:14:14,583 --> 05:14:17,052 NOT BRUT FORCE SIMULATIONS 6552 05:14:17,052 --> 05:14:18,587 WATCHING BINDS FOLD AND UNFOLD 6553 05:14:18,587 --> 05:14:21,156 AND THE RESEARCH GROUP HAS DONE 6554 05:14:21,156 --> 05:14:22,557 AN EXCELLENT JOB BUILDING A 6555 05:14:22,557 --> 05:14:23,825 COMPUTER TO RUN SIMULATIONS AT 6556 05:14:23,825 --> 05:14:25,827 THE SCALE AND MOST DON'T HAVE 6557 05:14:25,827 --> 05:14:26,895 THAT KIND OF COMPUTE CAPABILITY 6558 05:14:26,895 --> 05:14:28,763 AND WE NEED TO THINK MORE 6559 05:14:28,763 --> 05:14:29,965 CAREFULLY ABOUT THE PROBLEM 6560 05:14:29,965 --> 05:14:31,633 WE'RE TRYING TO SOLVE AND IN THE 6561 05:14:31,633 --> 05:14:34,569 CONTEXT OF OUR WORK AND DRUG 6562 05:14:34,569 --> 05:14:37,239 DISCOVERY THAT COMES DOWN TO 6563 05:14:37,239 --> 05:14:41,309 SAMPLING OR VARIOUS SAMPLING 6564 05:14:41,309 --> 05:14:43,478 PROTOCOLS TO INVESTIGATE 6565 05:14:43,478 --> 05:14:45,146 SPECIFIC DEGREES OF FREEDOM IN 6566 05:14:45,146 --> 05:14:50,619 THE CONTEXT OF WHAT MATTERS IN 6567 05:14:50,619 --> 05:14:52,787 BIOLOGY AND THIS PARTICULAR WORK 6568 05:14:52,787 --> 05:14:54,489 HERE DESCRIBE SOME OF THE TOOLS 6569 05:14:54,489 --> 05:14:55,056 WE'VE DEVELOPED ALONG THAT 6570 05:14:55,056 --> 05:14:58,560 PROFESSION. 6571 05:14:58,560 --> 05:15:03,164 THE COMPUTATION IS NOT ENOUGH. 6572 05:15:03,164 --> 05:15:04,599 WE CAN'T SIMULATE EVERYTHING AT 6573 05:15:04,599 --> 05:15:06,268 TIME SCALE AND RESOLUTION WE 6574 05:15:06,268 --> 05:15:06,568 NEED. 6575 05:15:06,568 --> 05:15:10,238 AND SO WE ALSO INTEGRATE BIO 6576 05:15:10,238 --> 05:15:11,940 PHYSICAL DATA AND DEVELOPED A 6577 05:15:11,940 --> 05:15:14,576 FRAMEWORK TO INTEGRATE DATA NOT 6578 05:15:14,576 --> 05:15:17,779 ATOMIC RESOLUTION BUT BRING IN 6579 05:15:17,779 --> 05:15:19,481 THAT ALLOWS US TO CREATE MORE 6580 05:15:19,481 --> 05:15:24,085 ACCURATE AND FASTER SIMULATIONS 6581 05:15:24,085 --> 05:15:26,988 AROUND WHAT MATTERS AND WHAT'S 6582 05:15:26,988 --> 05:15:29,090 HAPPENING IN THE EXPERIMENT. 6583 05:15:29,090 --> 05:15:30,592 FOR EXAMPLE MASS SPEC IS A 6584 05:15:30,592 --> 05:15:32,093 POWERFUL TOOL TO UNDERSTAND WHAT 6585 05:15:32,093 --> 05:15:36,131 PARTS OF A PROTEIN ARE GETTING 6586 05:15:36,131 --> 05:15:39,668 BLOCKED FROM SOLVENT UPON LIGAND 6587 05:15:39,668 --> 05:15:41,102 BINDING AND DOESN'T GIVE YOU 6588 05:15:41,102 --> 05:15:43,939 INFORMATION WHERE THE ATOMS ARE 6589 05:15:43,939 --> 05:15:46,574 JUST WHICH RESIDUES ARE BLOCKED 6590 05:15:46,574 --> 05:15:49,644 AND INTEGRATE THAT AS AN 6591 05:15:49,644 --> 05:15:51,479 ADDITIONAL FACTOR AND THE MD 6592 05:15:51,479 --> 05:15:52,881 SIMULATIONS AND THAT ALLOWS US 6593 05:15:52,881 --> 05:15:55,283 TO GET SIMULATIONS THAT ARE 6594 05:15:55,283 --> 05:15:56,251 CONSISTENT WITH EXPERIMENT. 6595 05:15:56,251 --> 05:15:58,920 THEY'RE NOT CONSTRAINED ONLY TO 6596 05:15:58,920 --> 05:16:02,557 THE EXPERIMENT BUT BIASSING THE 6597 05:16:02,557 --> 05:16:11,666 SIMULATIONS IN AN INTELLIGENT 6598 05:16:11,666 --> 05:16:13,668 WAY AND WE LOOKED AT MOLECULES 6599 05:16:13,668 --> 05:16:15,103 BRINGING TWO PROTEINS TOGETHER 6600 05:16:15,103 --> 05:16:17,305 AND THESE TEND TO BE PRETTY 6601 05:16:17,305 --> 05:16:19,841 FLEXIBLE AND WERE ABLE TO USE 6602 05:16:19,841 --> 05:16:23,445 THIS APPROACH WITH DYNAMICS AND 6603 05:16:23,445 --> 05:16:25,613 MASS SPEC TO GET ACCURATE 6604 05:16:25,613 --> 05:16:26,581 MOLECULAR REPRESENTATION 6605 05:16:26,581 --> 05:16:28,316 ESSENTIALLY ATOMIC CRYSTAL 6606 05:16:28,316 --> 05:16:31,052 RESOLUTION WITH FULL DYNAMICS 6607 05:16:31,052 --> 05:16:34,389 AND ABLE TO EXPLAIN PREVIOUSLY 6608 05:16:34,389 --> 05:16:34,923 UNEXPLAINABLE DATA ABOUT 6609 05:16:34,923 --> 05:16:37,492 SPECIFIC LINKERS AND HOW THOSE 6610 05:16:37,492 --> 05:16:38,560 ARE INFLUENTIAL IN THE 6611 05:16:38,560 --> 05:16:40,528 DEGRADATION PROFESSION. 6612 05:16:40,528 --> 05:16:41,029 -- PROCESS. 6613 05:16:41,029 --> 05:16:42,163 IT'S IMPORTANT AGAIN TO BRING 6614 05:16:42,163 --> 05:16:45,033 THIS BACK TO PEOPLE. 6615 05:16:45,033 --> 05:16:50,572 WE DEVELOPED AN INTEGRATIVE 6616 05:16:50,572 --> 05:16:52,707 DESIGN PROCESS WHERE WE START 6617 05:16:52,707 --> 05:16:57,145 AND END WITH HUMANS LOOKING AT 6618 05:16:57,145 --> 05:17:00,415 THE FREE ENERGY DECOMPOSITION 6619 05:17:00,415 --> 05:17:03,651 WORK I SHOWED AND OTHERS AND 6620 05:17:03,651 --> 05:17:07,122 ENUMERATE LARGE SPACES AND 6621 05:17:07,122 --> 05:17:09,157 WRITING OUT CHEMICAL REACTIONS 6622 05:17:09,157 --> 05:17:12,127 WITH KNOWN BUILDING BLOCKS TO 6623 05:17:12,127 --> 05:17:14,562 EXPLORE CHEMICAL SPACE AND LOOK 6624 05:17:14,562 --> 05:17:16,831 AT PREDICTIONS AND GETTING INTO 6625 05:17:16,831 --> 05:17:18,533 MOLECULAR DYNAMICS AND FREE 6626 05:17:18,533 --> 05:17:20,335 ENERGY CALCULATIONS AND ALL THIS 6627 05:17:20,335 --> 05:17:22,237 BEHIND THE SCENES GETS PRESENTED 6628 05:17:22,237 --> 05:17:25,940 TO PEOPLE ON THE TEAM THROUGH 6629 05:17:25,940 --> 05:17:26,574 THE DATA PLATFORM WE'VE 6630 05:17:26,574 --> 05:17:28,109 DEVELOPED SPECIFICALLY FOR THE 6631 05:17:28,109 --> 05:17:30,578 PURPOSE OF INTEGRATING HUMANS 6632 05:17:30,578 --> 05:17:33,148 AND COMPUTERS IN A WAY I THING 6633 05:17:33,148 --> 05:17:34,749 IS QUITE NOVEL AND USEFUL SO THE 6634 05:17:34,749 --> 05:17:36,317 RESULTS THAT COME OUT OF HERE 6635 05:17:36,317 --> 05:17:37,952 EVERYBODY ON THE TEAM GETS A 6636 05:17:37,952 --> 05:17:39,888 LINK AND CAN GO IN THE INTERFACE 6637 05:17:39,888 --> 05:17:42,590 AND CAN LOOK AT THE 6638 05:17:42,590 --> 05:17:43,858 COMPUTATIONAL DATA IN MANY 6639 05:17:43,858 --> 05:17:44,993 REPRESENTATIONS AND MAKE 6640 05:17:44,993 --> 05:17:46,394 DECISIONS AND PROVIDE OPINIONS 6641 05:17:46,394 --> 05:17:49,497 IT'S ALL COLLECTED SO WE'RE 6642 05:17:49,497 --> 05:17:51,166 DECIDING WHICH MOLECULES IT MAKE 6643 05:17:51,166 --> 05:17:52,500 AND CAPTURED EVERYBODY'S 6644 05:17:52,500 --> 05:17:55,770 FEEDBACK FROM BIOLOGY, 6645 05:17:55,770 --> 05:17:57,138 CHEMISTRY, BIO PHYSICS AND 6646 05:17:57,138 --> 05:17:57,439 COMPUTATION. 6647 05:17:57,439 --> 05:18:06,314 WE CAN'T CO -- DO THIS ALL 6648 05:18:06,314 --> 05:18:06,815 ALONE. 6649 05:18:06,815 --> 05:18:08,850 WE HAVE GREAT PEOPLE THAT HAVE 6650 05:18:08,850 --> 05:18:09,818 BEEN INVOLVED AND A LOT OF 6651 05:18:09,818 --> 05:18:13,221 THANKS TO THESE FOLKS AND THE 6652 05:18:13,221 --> 05:18:14,422 WORK THEY'RE DOING AS WELL AS 6653 05:18:14,422 --> 05:18:15,423 OTHER OPEN SCIENCE CONTRIBUTORS 6654 05:18:15,423 --> 05:18:18,393 NOT OUT THERE AND WE'RE ALSO 6655 05:18:18,393 --> 05:18:20,962 DOING OUR BIT ON THAT SIDE AND 6656 05:18:20,962 --> 05:18:24,833 DEVELOPING A NEW SIMULATION 6657 05:18:24,833 --> 05:18:27,769 ENGINE MADE TO BE MORE 6658 05:18:27,769 --> 05:18:32,440 PERFORMANT ON MODERN SIMULATIONS 6659 05:18:32,440 --> 05:18:34,042 WITH 50,000 ATOM DOESN'T 6660 05:18:34,042 --> 05:18:36,111 SATURATE A CPU AND THESE GROWING 6661 05:18:36,111 --> 05:18:39,747 SO FAST WE HAVE TO THINK MORE 6662 05:18:39,747 --> 05:18:42,584 CREATIVELY HOW TO LEVERAGE THEM 6663 05:18:42,584 --> 05:18:53,128 MAX -- MAXIMAL AND THIS IS OPEN 6664 05:18:54,629 --> 05:18:56,131 SOURCE UNDER THE MIT LICENSE. 6665 05:18:56,131 --> 05:18:57,966 SO WITH THAT AS BACKGROUND I 6666 05:18:57,966 --> 05:19:01,636 WANT TO JUMP INTO THE STORY FROM 6667 05:19:01,636 --> 05:19:03,138 OUR SILICON THERAPEUTICS DAYS 6668 05:19:03,138 --> 05:19:05,807 THE DESIGN OF A SYSTEMICALLY 6669 05:19:05,807 --> 05:19:10,578 AVAILABLE SMALL MOLECULE STING 6670 05:19:10,578 --> 05:19:11,646 AGONIST. 6671 05:19:11,646 --> 05:19:13,882 SO STING INTERFERON JOINS IS A 6672 05:19:13,882 --> 05:19:17,118 REGULATOR IN THE IMMUNE SYSTEM 6673 05:19:17,118 --> 05:19:17,752 AND IN RED. 6674 05:19:17,752 --> 05:19:21,956 IT GOES UNDER RAY LARGE SCALE 6675 05:19:21,956 --> 05:19:22,624 CONFIRMATIONAL CHANGE FROM THE 6676 05:19:22,624 --> 05:19:24,893 BLUE TO RED STRUCTURE WHEN I 6677 05:19:24,893 --> 05:19:29,664 BINDS AND THIS IS WHAT ACTIVATES 6678 05:19:29,664 --> 05:19:34,202 THE INNATE IMMUNE PATHWAY AND WE 6679 05:19:34,202 --> 05:19:38,573 DEVELOPED AN ANALOG TO DESIGN A 6680 05:19:38,573 --> 05:19:40,642 MOLECULE THAT'S SYSTEMICALLY 6681 05:19:40,642 --> 05:19:40,909 AVAILABLE. 6682 05:19:40,909 --> 05:19:45,146 WE HAD A NUMBER OF COMPANIES AND 6683 05:19:45,146 --> 05:19:49,050 TOOK NUCLEOTIDE AND IMPROVED THE 6684 05:19:49,050 --> 05:19:51,719 PROPERTIES AND THEY'RE LARGE AND 6685 05:19:51,719 --> 05:19:54,289 POLAR AND GET CLEARED OUT OF THE 6686 05:19:54,289 --> 05:19:59,594 BODY QUICKLY AND NON-DRUG LIKE 6687 05:19:59,594 --> 05:20:02,530 AND THESE ONLY SLIGHTLY IMPROVED 6688 05:20:02,530 --> 05:20:04,999 THE PROPERTIES BUT NOTHING TO BE 6689 05:20:04,999 --> 05:20:07,101 SYSTEMICALLY DELIVERED AND OUR 6690 05:20:07,101 --> 05:20:08,703 FEELING WAS A TRUE SMALL 6691 05:20:08,703 --> 05:20:11,773 MOLECULE COULD MAKE AN IMPACT 6692 05:20:11,773 --> 05:20:12,106 HERE. 6693 05:20:12,106 --> 05:20:15,176 I'LL SHOW A QUICK VIGNETTE AND 6694 05:20:15,176 --> 05:20:17,745 ANIMATION THAT COMBINES 6695 05:20:17,745 --> 05:20:18,813 DIFFERENT PARTS AND BREAK IT 6696 05:20:18,813 --> 05:20:23,518 DOWN TO THE INDIVIDUAL PIECES. 6697 05:20:23,518 --> 05:20:25,720 FIRST WE'RE RUNNING AT 6698 05:20:25,720 --> 05:20:27,188 STIMULATION AND I'LL PEEL BACK 6699 05:20:27,188 --> 05:20:31,426 THE WATER LOOKING AT THE STING 6700 05:20:31,426 --> 05:20:31,659 PROTEIN. 6701 05:20:31,659 --> 05:20:33,161 WE'LL SEE THE CONFIRMATIONAL 6702 05:20:33,161 --> 05:20:36,664 CHANGE FROM THE ACTIVE OPEN 6703 05:20:36,664 --> 05:20:38,733 CONFIRMATION TO WHEN A 6704 05:20:38,733 --> 05:20:39,901 NUCLEOTIDE IS BOUND FROM 6705 05:20:39,901 --> 05:20:40,668 INACTIVE TO ACTIVE. 6706 05:20:40,668 --> 05:20:43,037 WE WANT TO HAVE MOLECULES THAT 6707 05:20:43,037 --> 05:20:45,106 BIND TO THE ACTIVE CONFIRMATION. 6708 05:20:45,106 --> 05:20:49,477 WE DID THIS BY DESIGNING A SMALL 6709 05:20:49,477 --> 05:20:51,512 MOLECULE THAT DIMERIZES ON 6710 05:20:51,512 --> 05:20:54,449 ITSELF IN THE CONTEXT OF THE 6711 05:20:54,449 --> 05:20:55,283 BINDING SITE. 6712 05:20:55,283 --> 05:20:57,151 IT'S WHY SO MANY UNIQUE TOOLS 6713 05:20:57,151 --> 05:20:58,553 NEEDED TO BE DEVELOPED. 6714 05:20:58,553 --> 05:21:01,389 THIS IS INITIALLY AN IDEA. 6715 05:21:01,389 --> 05:21:03,758 THEY DIDN'T HAVE DETECTABLE 6716 05:21:03,758 --> 05:21:07,195 BINDING AFFINITY BUT BY 6717 05:21:07,195 --> 05:21:08,730 ANALYZING HIGH ENERGY WATER WE 6718 05:21:08,730 --> 05:21:12,834 WERE ABLE TO BUILD OUT INITIAL 6719 05:21:12,834 --> 05:21:13,401 POTENCY. 6720 05:21:13,401 --> 05:21:16,704 WE THEN STACK THE MOLECULE ON 6721 05:21:16,704 --> 05:21:17,071 ITSELF. 6722 05:21:17,071 --> 05:21:19,540 AGAIN IT'S A NOVEL MECHANISM. 6723 05:21:19,540 --> 05:21:23,177 AND WE WERE ABLE TO GET CRYSTAL 6724 05:21:23,177 --> 05:21:26,047 STRUCTURES AND GO THROUGH 6725 05:21:26,047 --> 05:21:29,050 QUANTUM MECHANICAL CALCULATION 6726 05:21:29,050 --> 05:21:33,421 AND GET INTO FREE BINDING ENERGY 6727 05:21:33,421 --> 05:21:35,089 CALCULATION USING WATER THERMAL 6728 05:21:35,089 --> 05:21:37,925 DYNAMICS TO GUIDE WHERE WE 6729 05:21:37,925 --> 05:21:41,329 WANTED TO GO AND SYNTHESIZABLE 6730 05:21:41,329 --> 05:21:44,766 LIBRARIES AND WE IDENTIFIED SOME 6731 05:21:44,766 --> 05:21:45,867 INTERESTING UNIQUE CONFIRMATION 6732 05:21:45,867 --> 05:21:49,370 IN ORANGE IS A CON FIRM OF THE 6733 05:21:49,370 --> 05:21:51,139 PROTEIN THAT LED TO A TIGHT 6734 05:21:51,139 --> 05:21:53,207 BINDING MOLECULE BUT DIDN'T HAVE 6735 05:21:53,207 --> 05:21:57,145 A STRONG PATHWAY STIMULATION. 6736 05:21:57,145 --> 05:22:01,582 WE HAD TO STEP BACK AND DESIGN 6737 05:22:01,582 --> 05:22:03,151 MOLECULES THAT BLOCK THE PROTEIN 6738 05:22:03,151 --> 05:22:05,219 INTO A HIGHLY ACTIVE CONFIRM AND 6739 05:22:05,219 --> 05:22:07,322 ONCE WE WERE ABLE TO IDENTIFY 6740 05:22:07,322 --> 05:22:09,190 THAT CHEMICAL SERIES WE THEN 6741 05:22:09,190 --> 05:22:10,458 FOCUSSED ON OPTIMIZING THE 6742 05:22:10,458 --> 05:22:12,493 PROPERTIES TO BRING THAT 6743 05:22:12,493 --> 05:22:14,062 MOLECULE FROM A LEAD TO A DRUG 6744 05:22:14,062 --> 05:22:16,164 AND THIS IS CURRENTLY IN THE 6745 05:22:16,164 --> 05:22:25,907 CLINIC. 6746 05:22:25,907 --> 05:22:30,411 WE STARTED FROM AN IDEA INTEREST 6747 05:22:30,411 --> 05:22:32,680 FROM THE MOUSE STEAM PROTEIN AND 6748 05:22:32,680 --> 05:22:35,350 LAD NO DETECTABLE BINDING 6749 05:22:35,350 --> 05:22:36,984 AFFINITY OR ACTIVITY TO HUMAN 6750 05:22:36,984 --> 05:22:37,852 STING. 6751 05:22:37,852 --> 05:22:39,153 WE STARTED DOING MODELLING AND 6752 05:22:39,153 --> 05:22:40,888 CAME UP WITH THE IDEA AND 6753 05:22:40,888 --> 05:22:42,523 OPTIMIZED THE MOLECULE. 6754 05:22:42,523 --> 05:22:45,159 WE WENT THROUGH VARIOUS STAGES 6755 05:22:45,159 --> 05:22:47,962 OF LOOKING AT THE WATER THERMAL 6756 05:22:47,962 --> 05:22:49,597 DYNAMICS AND FREE ENERGY 6757 05:22:49,597 --> 05:22:50,498 CALCULATION. 6758 05:22:50,498 --> 05:22:52,333 ALL REQUIRED CUSTOMIZED TOOLS TO 6759 05:22:52,333 --> 05:22:52,834 BE DEVELOPED. 6760 05:22:52,834 --> 05:22:54,102 THERE'S NOTHING OUT OF THE BOX 6761 05:22:54,102 --> 05:22:56,504 THAT ALLOWS YOU TO LOOK AT 6762 05:22:56,504 --> 05:22:57,705 MOLECULES BINDING AS A DIMER TO 6763 05:22:57,705 --> 05:22:59,807 EACH OTHER AND INTERACTING WITH 6764 05:22:59,807 --> 05:23:02,510 THE PROTEIN. 6765 05:23:02,510 --> 05:23:05,012 A VALUE OF OPEN TOOLS, 6766 05:23:05,012 --> 05:23:07,515 CUSTOMIZABLE TO HAVE A TEAM TO 6767 05:23:07,515 --> 05:23:09,751 GET IN THERE AND MAKE 6768 05:23:09,751 --> 05:23:14,322 MODIFICATIONS AS NEEDED. 6769 05:23:14,322 --> 05:23:18,559 THE FIRST WAS THE WATER THERMAL 6770 05:23:18,559 --> 05:23:22,530 DYNAMICS. 6771 05:23:22,530 --> 05:23:23,598 GREAT TOOLS OUT THERE AND WE 6772 05:23:23,598 --> 05:23:26,567 WORKED WITH THE DEVELOPER OF 6773 05:23:26,567 --> 05:23:28,202 WATER MAP AND THERE'S A GREAT 6774 05:23:28,202 --> 05:23:31,539 TOOL OUT THERE THAT GIVES YOU 6775 05:23:31,539 --> 05:23:34,409 PICTURES LIKE THIS AND WITH THE 6776 05:23:34,409 --> 05:23:35,410 INITIAL HYPOTHESIS OF WHERE THIS 6777 05:23:35,410 --> 05:23:37,478 WAS IT'S EASY TO LOOK AT A 6778 05:23:37,478 --> 05:23:39,647 PICTURE AND SAY IF WE COULD JUST 6779 05:23:39,647 --> 05:23:42,116 GET OUT THERE WE COULD GAIN 6780 05:23:42,116 --> 05:23:44,085 POTENCY AND THAT ENDED UP 6781 05:23:44,085 --> 05:23:46,387 WORKING AT THE HIGHEST ENERGY 6782 05:23:46,387 --> 05:23:47,088 WATER PATCHES. 6783 05:23:47,088 --> 05:23:50,024 AND ONCE BEGOT THE INITIAL 6784 05:23:50,024 --> 05:23:52,393 DETECTABLE BINDERS WE GOT INTO 6785 05:23:52,393 --> 05:23:53,895 THE STACKING INTERACTION. 6786 05:23:53,895 --> 05:23:58,833 EVEN WITH SMALL HALOGEN AND 1 TO 6787 05:23:58,833 --> 05:24:00,601 2 ATOM CHANGES THERE'S 6788 05:24:00,601 --> 05:24:03,137 COMBINATIONS ONE CAN DO WITH THE 6789 05:24:03,137 --> 05:24:04,372 EIGHT POTENTIAL SITES OF 6790 05:24:04,372 --> 05:24:04,639 CHEMISTRY. 6791 05:24:04,639 --> 05:24:11,712 WE ENUMERATED THE IDEAS AND RAN 6792 05:24:11,712 --> 05:24:16,451 THIS FUNCTIONAL SYMMETRY THEORY 6793 05:24:16,451 --> 05:24:18,820 AND WORKED WITH TODD MARTINEZ 6794 05:24:18,820 --> 05:24:22,557 WITH STANFORD THE DEVELOPER OF A 6795 05:24:22,557 --> 05:24:24,625 FAST QUANTUM MECHANICAL TOOL 6796 05:24:24,625 --> 05:24:29,564 THAT RUNS ON GPUs AND IN 6797 05:24:29,564 --> 05:24:31,532 COLLABORATION WERE ABLE TO VFRL 6798 05:24:31,532 --> 05:24:33,701 THE TOOL IN REAL TIME FROM THE 6799 05:24:33,701 --> 05:24:34,402 IDENTIFICATION OF THE NEED AND 6800 05:24:34,402 --> 05:24:37,405 WE HAD A TOOL TO COMPUTE THE 6801 05:24:37,405 --> 05:24:39,707 STRENGTH OF THE STACKING 6802 05:24:39,707 --> 05:24:40,842 INTERACTION AND THAT CORRELATES 6803 05:24:40,842 --> 05:24:42,677 WELL WITH THE ACTIVITY WE WERE 6804 05:24:42,677 --> 05:24:44,745 LOOKING TO BUILD AND WERE ABLE 6805 05:24:44,745 --> 05:24:48,783 TO EXPLORE THOUSANDS OF 6806 05:24:48,783 --> 05:24:49,317 MOLECULES VIRTUALLY AND 6807 05:24:49,317 --> 05:24:52,753 SYNTHESIZE ONLY 10 TO COME UP 6808 05:24:52,753 --> 05:24:57,158 WITH A MOLECULE MORE POTENT THAN 6809 05:24:57,158 --> 05:24:59,460 WHERE WE WERE AND IMPROVING 6810 05:24:59,460 --> 05:25:00,995 PROPERTIES WHILE MAINTAINING AND 6811 05:25:00,995 --> 05:25:02,763 IMPROVING BINDING AFFINITY. 6812 05:25:02,763 --> 05:25:04,799 THIS IS A DISTRIBUTION YOU'RE 6813 05:25:04,799 --> 05:25:06,567 LOOKING AT ON THE RIGHT WHICH IS 6814 05:25:06,567 --> 05:25:09,837 ALL THE 200 OR SO COMPOUNDS OVER 6815 05:25:09,837 --> 05:25:12,340 THE COURSE OF THE PROJECT. 6816 05:25:12,340 --> 05:25:15,409 I LIKE TO SHOW THE CORRELATION 6817 05:25:15,409 --> 05:25:17,812 PLOTS AND RMS ERRORS AND THIS IS 6818 05:25:17,812 --> 05:25:21,182 WHAT YOU'D WANT TO SEE AT VEGAS 6819 05:25:21,182 --> 05:25:22,683 MAKING A BET ON SOMETHING. 6820 05:25:22,683 --> 05:25:27,154 THERE'S A PREDICTIVE VALUE AND 6821 05:25:27,154 --> 05:25:28,789 MOLECULES WE PREDICT IN THE 6822 05:25:28,789 --> 05:25:30,558 FLUID DISTRIBUTION AND END UP 6823 05:25:30,558 --> 05:25:32,393 BEING BETTER AND SELECT FROM A 6824 05:25:32,393 --> 05:25:34,562 POOL OF MOLECULES WE'D PREFER 6825 05:25:34,562 --> 05:25:38,566 THIS BLUE DISTRIBUTE UKS THAN 6826 05:25:38,566 --> 05:25:44,639 AVERAGE DISTRIBUTION AND 6827 05:25:44,639 --> 05:25:48,342 MOLECULES WE PREDICT TO BE BAD 6828 05:25:48,342 --> 05:25:51,979 BUT ENDED UP SYNTHESIZING AND 6829 05:25:51,979 --> 05:25:52,947 FALSE NEGATIVES. 6830 05:25:52,947 --> 05:25:54,282 WE WERE PROBABLY TRYING TO 6831 05:25:54,282 --> 05:25:57,151 IMPROVE OTHER PROPERTIES BEYOND 6832 05:25:57,151 --> 05:26:01,455 THE BINDING AND MOLECULE WE 6833 05:26:01,455 --> 05:26:03,524 PREDICTED TO BE FALSE POSITIVES 6834 05:26:03,524 --> 05:26:05,893 BUT IN GENERAL HIGHLY PREDICTIVE 6835 05:26:05,893 --> 05:26:07,128 AND USEFUL AND SPECIFIC EXAMPLES 6836 05:26:07,128 --> 05:26:10,565 OF THE TYPES OF PERTURBATIONS WE 6837 05:26:10,565 --> 05:26:14,902 WERE MAKING AND THE EXPERIMENTAL 6838 05:26:14,902 --> 05:26:15,436 BINDING ENERGY VERSUS OUR 6839 05:26:15,436 --> 05:26:23,444 ENERGIES. 6840 05:26:23,444 --> 05:26:25,580 FINDINGS AREN'T ALWAYS WHAT 6841 05:26:25,580 --> 05:26:25,913 MATTERS. 6842 05:26:25,913 --> 05:26:27,815 WE WERE LOOKING FOR 6843 05:26:27,815 --> 05:26:29,817 STABILIZATION OF A GOOD 6844 05:26:29,817 --> 05:26:30,484 CONFIRMATION AND EXPERIMENTALLY 6845 05:26:30,484 --> 05:26:38,559 IN THE LAB BINDING AFFINITY 6846 05:26:38,559 --> 05:26:41,128 DIDN'T CORRELATE WITH CELL 6847 05:26:41,128 --> 05:26:45,833 ACTIVITY AND CAME UP WITH A 6848 05:26:45,833 --> 05:26:46,567 STABILIZATION ASSAY THAT 6849 05:26:46,567 --> 05:26:50,538 CORRELATED WITH WHAT WE CARED 6850 05:26:50,538 --> 05:26:52,406 ABOUT CELLULAR ACTIVITY. 6851 05:26:52,406 --> 05:26:53,941 AND TO MENTION IT'S IMPORTANT TO 6852 05:26:53,941 --> 05:26:55,109 KNOW WHICH YOU'RE GOING AFTER 6853 05:26:55,109 --> 05:26:56,844 AND WHAT YOUR VARIOUS MODELS ARE 6854 05:26:56,844 --> 05:26:58,546 GOING TO BE THROUGHOUT THE 6855 05:26:58,546 --> 05:26:59,280 PROCESS. 6856 05:26:59,280 --> 05:27:00,781 ONE PROBLEM WITH THE EARLIER 6857 05:27:00,781 --> 05:27:04,919 MOLECULE IS YOU COULD SEE IT 6858 05:27:04,919 --> 05:27:06,153 STABILIZES THE MOUSE VERSION BUT 6859 05:27:06,153 --> 05:27:10,558 HAD NO ACTIVITY ON HUMAN. 6860 05:27:10,558 --> 05:27:17,765 IN THE PROCESS WE LOOKED AT THE 6861 05:27:17,765 --> 05:27:21,202 HUMAN WILD TYPE ALLELE AND THE 6862 05:27:21,202 --> 05:27:23,037 PRECLINICAL MODELS. 6863 05:27:23,037 --> 05:27:26,040 WE WANT TO MAKE SURE THEY'LL 6864 05:27:26,040 --> 05:27:30,478 WORK IN MOUSE MODELS. 6865 05:27:30,478 --> 05:27:32,680 THE REASON IT MATTERS IS THE 6866 05:27:32,680 --> 05:27:36,350 MAIN CONCERN IS ON TARGET TALKS 6867 05:27:36,350 --> 05:27:36,884 CYTOKINE STORM. 6868 05:27:36,884 --> 05:27:39,086 WE WANTED TO MAKE SURE IT WAS 6869 05:27:39,086 --> 05:27:40,888 BINDING AND DOING WHAT IT WAS 6870 05:27:40,888 --> 05:27:43,024 DOING ACROSS THE RELEVANT 6871 05:27:43,024 --> 05:27:43,290 SPECIES. 6872 05:27:43,290 --> 05:27:45,559 THROUGH THE PROJECT WE WERE ABLE 6873 05:27:45,559 --> 05:27:49,597 TO DESIGN A POTENT AND SELECTIVE 6874 05:27:49,597 --> 05:27:51,732 SMALL MOLECULE AGONIST THAT 6875 05:27:51,732 --> 05:27:57,204 WORKED ON HHQ AND OTHER VARIANTS 6876 05:27:57,204 --> 05:28:05,513 MORE POTENT AND THE ANALOGS AND 6877 05:28:05,513 --> 05:28:08,883 SINGLE DOSE RESULTED IN A 6878 05:28:08,883 --> 05:28:10,351 MEDIATED MECHANISM. 6879 05:28:10,351 --> 05:28:18,993 THE GOAL ISN'T TO INHIBIT THE 6880 05:28:18,993 --> 05:28:24,432 PROTEIN BUT TURN THE BODY ON AND 6881 05:28:24,432 --> 05:28:26,033 DO THE REST AND YOU SEE THE 6882 05:28:26,033 --> 05:28:28,436 TREATED AND UNTREATED POPULATION 6883 05:28:28,436 --> 05:28:30,571 AND SINGLE DOSE COMPLETED IN 6884 05:28:30,571 --> 05:28:34,208 COMPLETE TUMOR ELIMINATION. 6885 05:28:34,208 --> 05:28:37,511 MORE INTERESTINGLY THE PERFECTLY 6886 05:28:37,511 --> 05:28:38,946 HEALTHY MICE WERE RE-INOCULATED 6887 05:28:38,946 --> 05:28:43,384 WITH THE TUMOR AND WHILE IT GREW 6888 05:28:43,384 --> 05:28:46,320 IN NAIVE MICE THE PREVIOUSLY 6889 05:28:46,320 --> 05:28:48,789 TREATED MICE WITHOUT ADDITIONAL 6890 05:28:48,789 --> 05:28:52,760 TREATMENT DIDN'T GROW. 6891 05:28:52,760 --> 05:28:54,261 THERE WAS INNATE LEARNING IN THE 6892 05:28:54,261 --> 05:28:54,495 PROCESS. 6893 05:28:54,495 --> 05:28:57,264 THIS IS A SUMMARY OF THE TOOLS I 6894 05:28:57,264 --> 05:29:00,434 COVERED AND USED BUT THE MAIN 6895 05:29:00,434 --> 05:29:01,836 CONCLUSION IS WE EXPLORED 6896 05:29:01,836 --> 05:29:05,506 MILLIONS VIRTUALLY AND APPLIED 6897 05:29:05,506 --> 05:29:09,076 CUSTOM TOOLS IN THE PROJECT 6898 05:29:09,076 --> 05:29:12,179 BECAUSE OF THE UNIQUE PROBLEMS 6899 05:29:12,179 --> 05:29:17,718 OF THE CHAIN AND BACK TO THE 6900 05:29:17,718 --> 05:29:18,552 CLINICAL MOLECULE. 6901 05:29:18,552 --> 05:29:19,920 DRUG DISCOVERY IS HARD FOR 6902 05:29:19,920 --> 05:29:22,022 DIFFERENT REASONS AND BLACK 6903 05:29:22,022 --> 05:29:24,558 BOXES ARE NOT THERE FOR DRUG 6904 05:29:24,558 --> 05:29:24,825 DISCOVERY. 6905 05:29:24,825 --> 05:29:26,560 WE NEED BESPOKE SOLUTIONS TO 6906 05:29:26,560 --> 05:29:28,429 SOLVE PROBLEMS AS THEY ARISE. 6907 05:29:28,429 --> 05:29:30,531 WE DEVELOPED A TOOL KIT APPROACH 6908 05:29:30,531 --> 05:29:32,366 TO APPLY THE RIGHT TOOL AT THE 6909 05:29:32,366 --> 05:29:34,568 RIGHT TIME. 6910 05:29:34,568 --> 05:29:36,270 COME PREPARED OR ANYTHING AND 6911 05:29:36,270 --> 05:29:37,805 EVERYTHING AND HAVE NOT JUST THE 6912 05:29:37,805 --> 05:29:38,606 TOOLS BUT PEOPLE TO DO 6913 05:29:38,606 --> 05:29:40,975 ADDITIONAL DEVELOPMENT WITH THE 6914 05:29:40,975 --> 05:29:42,977 TOOLS AS NECESSARY. 6915 05:29:42,977 --> 05:29:44,712 THE FUTURE HOLDS GREAT PROMISE 6916 05:29:44,712 --> 05:29:46,547 FOR A.I. 6917 05:29:46,547 --> 05:29:47,915 THERE'S GREAT STUFF COMING OUT 6918 05:29:47,915 --> 05:29:49,450 BUT WE LOVE IN THE PRESENT AND 6919 05:29:49,450 --> 05:29:52,853 HAVE TO USE WHAT'S THERE. 6920 05:29:52,853 --> 05:29:54,588 WE'RE INVESTING IN BOTH 6921 05:29:54,588 --> 05:29:56,524 DIRECTIONS AND HAPPY WITH THE 6922 05:29:56,524 --> 05:30:00,995 DIRECTION THINGS ARE MOVING IN 6923 05:30:00,995 --> 05:30:07,601 THE FIELD. 6924 05:30:07,601 --> 05:30:09,236 VERY IMPORTANTLY COMPUTERS ALONE 6925 05:30:09,236 --> 05:30:12,406 CAN'T INVOLVE THE PROBLEM. 6926 05:30:12,406 --> 05:30:15,242 IT'S THE INTEGRATION OF THE LAB 6927 05:30:15,242 --> 05:30:16,977 AND IN SILICO APPROACHES AND 6928 05:30:16,977 --> 05:30:20,247 APPROPRIATE PEOPLE TO MAKE THE 6929 05:30:20,247 --> 05:30:21,248 DECISIONS. 6930 05:30:21,248 --> 05:30:24,018 I'D LIKE TO THANK EVERYBODY AND 6931 05:30:24,018 --> 05:30:25,519 HAPPY TO ANSWER QUESTIONS. 6932 05:30:25,519 --> 05:30:25,986 >> THANK YOU FOR THE 6933 05:30:25,986 --> 05:30:27,254 PRESENTATION. 6934 05:30:27,254 --> 05:30:37,765 WE'LL TAKE A COUPLE QUESTIONS. 6935 05:30:40,768 --> 05:30:42,503 WHAT MADE THE DIMERIZATION AND 6936 05:30:42,503 --> 05:30:50,711 STRUCTURE PROCESS? 6937 05:30:50,711 --> 05:30:55,649 >> AS THE THROUGH STACKING 6938 05:30:55,649 --> 05:30:56,951 INTERACTION AND THROUGH TO 6939 05:30:56,951 --> 05:31:02,556 ENTROPY AND CONCENTRATIONS IT'S 6940 05:31:02,556 --> 05:31:03,657 NOT OBSERVED. 6941 05:31:03,657 --> 05:31:08,829 THERE'S ANOTHER GROUP THAT 6942 05:31:08,829 --> 05:31:12,032 DEMONSTRATED SOME WEAK 6943 05:31:12,032 --> 05:31:13,500 AGGREGATION AND THE COMBINATION 6944 05:31:13,500 --> 05:31:17,438 OF THE BOUNDING SITE AND FIT 6945 05:31:17,438 --> 05:31:20,341 COUPLED WITH THAT INTERACTION 6946 05:31:20,341 --> 05:31:21,242 DRIVING THINGS TOWARDS THAT 6947 05:31:21,242 --> 05:31:23,777 THERMAL DYNAMIC STATE AND 6948 05:31:23,777 --> 05:31:26,213 PROTEIN PLUS THE MOLECULES 6949 05:31:26,213 --> 05:31:26,480 TOGETHER. 6950 05:31:26,480 --> 05:31:27,848 LIKELY THERE'S FAVORABLE 6951 05:31:27,848 --> 05:31:30,517 INTERACTION AND SOLUTION. 6952 05:31:30,517 --> 05:31:34,555 WE'RE MEASURING ON THUR -- THUR 6953 05:31:34,555 --> 05:31:38,492 OWN WHAT IS THE FAVORABLE 6954 05:31:38,492 --> 05:31:40,094 INTERACTION AND CONCENTRATION 6955 05:31:40,094 --> 05:31:41,729 AND SOLUTION WITH ENTROPY YOU 6956 05:31:41,729 --> 05:31:44,231 WON'T SEE THE DIMER. 6957 05:31:44,231 --> 05:31:46,567 THE MOLECULES LOOK AND BEHAVE 6958 05:31:46,567 --> 05:31:51,005 LIKE MONODISPERSED TRUE SMALL 6959 05:31:51,005 --> 05:31:54,475 MOLECULES WITH SOLUBILITY AND 6960 05:31:54,475 --> 05:31:57,244 PERMEABILITY OF A SMALL MOLECULE 6961 05:31:57,244 --> 05:32:01,215 AND WITH A POPULATION DIMERIZED 6962 05:32:01,215 --> 05:32:04,285 AND STABILIZING THE ACTIVE 6963 05:32:04,285 --> 05:32:05,886 CONFIRMATION OF STING. 6964 05:32:05,886 --> 05:32:10,491 >> YOU ARE USING YOUR OWN ENGINE 6965 05:32:10,491 --> 05:32:13,027 AS FAR AS I UNDERSTOOD. 6966 05:32:13,027 --> 05:32:14,561 WHAT FORCE FIELD DO YOU USE HAVE 6967 05:32:14,561 --> 05:32:19,199 YOU TRIED OR PLAN TO TRY 6968 05:32:19,199 --> 05:32:21,902 CREATING YOUR OWN ML BASED FORCE 6969 05:32:21,902 --> 05:32:25,906 FIELD ALONG THE LINES JOHN 6970 05:32:25,906 --> 05:32:27,241 OUTLINED EARLIER. 6971 05:32:27,241 --> 05:32:28,509 >> WE ARE VERY MUCH KEEPING WITH 6972 05:32:28,509 --> 05:32:29,944 THE FIELD THROUGH OUR 6973 05:32:29,944 --> 05:32:31,845 DEVELOPMENT AND WORKING WITH 6974 05:32:31,845 --> 05:32:34,114 WHAT'S OUT THERE. 6975 05:32:34,114 --> 05:32:37,318 INTERNALLY WE DEVELOPED AN 6976 05:32:37,318 --> 05:32:40,254 INFRASTRUCTURE THAT DOES BESPOKE 6977 05:32:40,254 --> 05:32:44,391 DIMERIZATION AND PORTION SCANS 6978 05:32:44,391 --> 05:32:46,560 FROM MECHANICAL FITTING AND THE 6979 05:32:46,560 --> 05:32:53,467 STATE OF APPROACH AND OPEN FORCE 6980 05:32:53,467 --> 05:32:55,169 FIELD WASN'T A THING WHEN WE 6981 05:32:55,169 --> 05:33:03,243 STARTED THIS AND WE NOW USE IT 6982 05:33:03,243 --> 05:33:04,044 QUITE REGULARLY AND THERE'S 6983 05:33:04,044 --> 05:33:08,816 PROMISE THERE. 6984 05:33:08,816 --> 05:33:09,249 >> THANK YOU. 6985 05:33:09,249 --> 05:33:14,355 WE'LL CONTINUE ONE THE TOPIC 6986 05:33:14,355 --> 05:33:18,559 TOMORROW WITH THE SPEAKERS. 6987 05:33:18,559 --> 05:33:21,729 AND I'LL HAND IT OVER TO SARINE 6988 05:33:21,729 --> 05:33:28,302 FOR FINAL COMMENTS. 6989 05:33:28,302 --> 05:33:31,372 >> THANK YOU SO MUCH AND THANK 6990 05:33:31,372 --> 05:33:33,874 YOU FOR THE SPEAKERS AS WELL AS 6991 05:33:33,874 --> 05:33:36,977 ALL THE SPEAKERS TODAY ALL 6992 05:33:36,977 --> 05:33:38,545 SESSION CHAIRS AND EVERYONE 6993 05:33:38,545 --> 05:33:39,713 ELSE. 6994 05:33:39,713 --> 05:33:42,516 IT WAS PHENOMENAL. 6995 05:33:42,516 --> 05:33:46,286 THIS WAS AN AWESOME DAY. 6996 05:33:46,286 --> 05:33:49,089 I'D LIKE TO SHARE THE AGENDA FOR 6997 05:33:49,089 --> 05:33:50,324 TOMORROW QUICKLY AGAIN TO REMIND 6998 05:33:50,324 --> 05:33:56,363 YOU ALL ABOUT THE EXCITING TALKS 6999 05:33:56,363 --> 05:33:57,431 WE'RE GOING HAVE TOMORROW. 7000 05:33:57,431 --> 05:33:58,565 PLEASE COME BACK TOMORROW. 7001 05:33:58,565 --> 05:34:01,802 WE HAVE EXCITING SESSIONS. 7002 05:34:01,802 --> 05:34:02,536 WE'RE GOING TO CONTINUE SESSION 7003 05:34:02,536 --> 05:34:05,539 3 ON APPLICATIONS OF METHOD 7004 05:34:05,539 --> 05:34:07,408 METHODOLOGIES AND DRUG 7005 05:34:07,408 --> 05:34:07,674 DISCOVERY. 7006 05:34:07,674 --> 05:34:11,345 WE'LL HAVE A SESSION FOCUSES ON 7007 05:34:11,345 --> 05:34:14,548 EMERGING TRENDS THAT WILL HAVE A 7008 05:34:14,548 --> 05:34:16,617 NEWLY MINTED NOBEL LAUREATE 7009 05:34:16,617 --> 05:34:19,219 SPEAK AT 2:00 IN THE AFTERNOON 7010 05:34:19,219 --> 05:34:23,490 AND THEN RIGHT AFTER WE'LL HAVE 7011 05:34:23,490 --> 05:34:24,892 A PANEL DISCUSSION THAT IS GOING 7012 05:34:24,892 --> 05:34:27,494 TO COVER BRIDGING THE GAPS IN 7013 05:34:27,494 --> 05:34:28,829 TRANSLATION WHERE WE'RE GOING TO 7014 05:34:28,829 --> 05:34:32,733 HAVE AWESOME DISCUSSIONS AND 7015 05:34:32,733 --> 05:34:35,102 HAVE A LOT OF TIME AFTER THE 7016 05:34:35,102 --> 05:34:39,773 SHORT TALKS OF THE PANELISTS TO 7017 05:34:39,773 --> 05:34:40,908 DISCUSS MORE. 7018 05:34:40,908 --> 05:34:42,643 IF WE DIDN'T HIT ANY OF YOUR 7019 05:34:42,643 --> 05:34:44,244 QUESTIONS OR MISSED ANYTHING 7020 05:34:44,244 --> 05:34:45,245 THIS IS THE TIME TO COME BACK 7021 05:34:45,245 --> 05:34:49,116 AND ASK YOUR QUESTIONS. 7022 05:34:49,116 --> 05:34:51,085 I'D ALSO LIKE TO THANK THE 7023 05:34:51,085 --> 05:34:51,351 ATTENDEES. 7024 05:34:51,351 --> 05:34:55,389 I KNOW IT'S LATE IN THE NIGHT OR 7025 05:34:55,389 --> 05:34:57,257 SUPER EARLY IN THE MORNING. 7026 05:34:57,257 --> 05:35:00,928 WE APPRECIATE YOU BEING HERE 7027 05:35:00,928 --> 05:35:02,963 WITH US AND YOUR EAGERNESS TO 7028 05:35:02,963 --> 05:35:03,230 LEARN. 7029 05:35:03,230 --> 05:35:09,236 I'D LIKE TO CLOSE TODAY'S 7030 05:35:09,236 --> 05:35:11,271 SESSION AND SEE YOU BACK 7031 05:35:11,271 --> 05:35:16,410 TOMORROW 11:00 EASTERN TIME. 7032 05:35:16,410 --> 05:35:18,712 THANK YOU EVERYONE. 7033 05:35:18,712 --> 05:35:29,022 BYE, GOOD EVENING.