>> CAN WE GET STARTED, PLEASE. THANK YOU VERY MUCH. WELCOME THIS MORNING, EVERYBODY MADE IT, YOU SURVIVED THE PARTY LAST NIGHT. IT WAS WONDERFUL. SO THANK YOU ALL FOR -- THAT WAS KIND OF AN INTERESTING BUS RIDE. IT WILL BE ONE OF THE MEMORABLE EVENTS OF THIS MEETING. WE HAVE GOT TOO MANY OF THOSE. THAT WILL BE MY NEXT THING. SEE IF WE CAN HAVE A MEMORABLE BUS RIDE BUT PROBABLY NO WAY. WE HAD A GOOD CHAD. GOOD NETWORKING. SO IT'S MY PLEASURE TO INTRODUCE BARBARA RAPP, DIRECTOR OF PLANNING FROM THE OFFICE OF THE DIRECTOR. SHE WILL BE TALKING TO US ABOUT AND PRESENTING THE RESULTS OR RESPONSES TO THE R-5 O RFI ON THE NLM WORKING GROUP. TURN IT OVER TO BARBARA. >> GOOD MORNING. IT'S TRULY A PLEASURE TO REPORT TO YOU ON THE REQUEST FOR INFORMATION THAT WAS ISSUED IN THE SPRING AS PART OF THE WORKING GROUP THAT DR. COLLINS CONVENED TO CONSIDER FUTURE DIRECTIONS. POURING OVER SOME OF THE EXTENSIVE COMMENTS SUBMITTED BY THE WIDE RANGE OF PEOPLE WHO USE OUR SERVICES REINFORCE WHAT A PRIVILEGE IT IS TO BE PART OF THE GREAT ORGANIZATION. TAKE A LOOK AT SOME OF THE SHORT STATEMENTS THAT CAME IN. I OFTEN FELT READING THESE I WAS IN THE MIDDLE OF THE WAR OF FAVORITES T. OVERALL SENTIMENT IS REFLECTED IN COMMENT AT THE TOP RIGHT. WE CAN'T OVERSTATE THE IMPORTANCE OF THE NLM RESOURCES. YES, THERE WERE SOME NEGATIVES, THERE WERE MANY SUGGESTIONS FOR IMPROVEMENTS AND SOME RECOMMENDATIONS FOR CHANGES IN DIRECTION. BUT THE INPUT WAS OVERWHELMINGLY SUPPORTED IN NATURE AND HIGHLY THOUGHTFUL AND DELIBERATIVE. THOUGH SOME WERE SHORT STATEMENTS SUCH AS THE ONES LISTED HERE, THE INPUT GENERALLY CONSISTED OF ONE OR TWO PARAGRAPHS WHICH COME ABOUT A PAGE AND A FEW THAT EXTENDED TO FOUR OR FIVE PAGES. IN MY TALK TODAY I'LL FIRST PROVIDE SUMMARY INFORMATION ABOUT THE RFI AND THE COMMENTS THAT WERE RECEIVED AND OUTLINE SOME OF THE MAIN THEMES FOR THE FUTURE AS A PLATFORM FOR DISCUSSION. EARLIER THIS YEAR FOR 30 DAYS FEBRUARY AND MARCH REQUEST FOR INFORMATION WAS ISSUED BY THE WORKING GROUP ON NLM. AS I MENTIONED CONVENED BY SUBCOMMITTEE OF THE ADVISORY COMMITTEE TO DIRECTOR CONVENED BY DR. COLLINS. INPUT WAS REQUESTED ON FIVE QUESTIONS. AND THERE WERE 649 RESPONDENTS. THE EXACT TEXT OF THE QUESTION IS INCLUDED UNDER TAB 7 IN YOUR BOARD BOOKS BUT THE BASIC STRUCTURE OF THE FIRST FOUR QUESTIONS WAS THEY ANDED THE CURRENT NLM ELEMENTS THAT WERE THE MOST OR LEAST VALUE TO FOUR GROUPS. RESEARCH COMMUNITY, HEALTH PROFESSIONALS PATIENTS IN THE PUBLIC AND THEY LUMP LIBRARIES PUBLISHERS ORGANIZATIONS AND DEVELOPER WHOSE USE OUR DATA TO PROVIDE COMPLIMENTARY SERVICES TOGETHER. THEN -- AND THEY ALSO ASK FOR EACH OF THOSE FUTURE CAPABILITIES NEEDED WITHIN THAT ARENA. FOR QUESTION FIVE THEY ASKED HOW NLM COULD BE BETTER POSITIONED TO ADDRESS THE BROADER AND GROWING CHALLENGES ASSOCIATED WITH THE AREAS LISTED HERE INFORMATICS, DATA SCIENCE, EHR, DIGITAL PUBLICATION. AND ANY OTHER EMERGING CHALLENGES THAT THEY CARE TO NOTE. THE CONTENT OF THE RESPONSES COVER THE BREADTH OF NLM MORE THAN 200 PRODUCTS FEATURES PROGRAMS NAMED AND THEY WERE CONTRIBUTED BY THE BROAD SPECTRUM OF COMMUNITIES WE SERVE. THIS TABLE LISTS THE ROLE OF THE RESPONDENTS WITH LIBRARIANS MAKING UP THE LARGEST SEGMENT, 38.7% FOLLOWED BY RESEARCHERS, EMERGENCY PERSONNEL, EDUCATORS HISTOSTORIANS. HEALTH PROFESSIONALS AND THOSE IN ENVIRONMENTAL TOXICOLOGY. FEW GROUPS WITH LESS THAN 5% REPRESENTATION OR PUBLIC HEALTH WORKERS, DISEASE ADVOCATES AND PUBLISHERS COMMUNITY. THEY CAME FROM A VARIETY OF INSTITUTION TYPES OF COURSE THE LARGEST BEING UNIVERSITY AT 24.3% FOLLOWED BY CAN'T REALLY SAY FOLLOWED BY BUT NEXT IS MEDICAL SCHOOL. I BROKE OUT MEDICAL SCHOOL FROM UNIVERSITY WHERE THAT WAS CLEAR FROM THE DEMOGRAPHIC INFORMATION PROVIDED AND THAT ALSO INCLUDES THE MEDICAL SCHOOL LIBRARIES FOLLOWED BY HOSPITALS PRIVATE COMPANIES STATE AND LOCAL GOVERNMENT GROUPS, PROFESSIONAL AND TRADE ASSOCIATIONS AND FEDERAL GOVERNMENT, HERE I ALSO BROKE OUT YOU CAN SEE BY THE ASTERICS THE MILITARY AND NIH SEPARATELY AND OTHER GROUPS AT THE LESS THAN 5% WERE CHARACTERS-23 COMMUNITY, -- K-IS THE COMMUNITY, PUBLIC LIBRARIES INTERNATIONAL GROUPS, REGIONAL MEDICAL LIBRARIES AND COMMUNITY COLLEGES. THE PROFESSIONAL ASSOCIATIONS THAT WERE REPRESENTED INCLUDE OUR KEY COLLABORATORS AMONG OTHERS, MLA, AAMC, ACME ASSOCIATION OF RESEARCH LIBRARIES AND THE AMERICAN ASSOCIATION FOR THE HISTORY OF MEDICINE. AND THIS IS A GOOD PLACE I THINK TO PAUSE FOR A SPECIAL THANK YOU TO THE BOARD FOR YOUR EFFORTS IN CONTACTING ORGANIZATIONS SUCH AS THESE AND ASKING FOR THEIR PARTICIPATION. THAT WAS A VERY IMPORTANT COMPONENT. IN TERMS OF RECEIVING THE COMMENTS THAT WE DID. SO OF THE TOTAL 649 RESPONDENTS THEY DIDN'T ALL RESPOND TO EACH QUESTION. THEY RESPONDED TO, ABOUT THREE COMMENTS PER RESPONDENT. SO THERE WERE 421 RESPONSES TO QUESTION ONE ON THE VALUE OF RESEARCHERS AND THEN YOU CAN SEE THE NUMBERS FOR THE OTHER FOUR, 20 FOR HEALTH PROFESSIONALS, 277 FOR PATIENTS AND GENERAL PUBLIC. AND THEN THE LIBRARIANS AND THE COMMENTS FOR THE FUTURE. THIS TOTALS UP TO JUST UNDER 1900, 1898. ABOUT 1900 INDIVIDUAL COMMENTS RECEIVED. THIS SET OF TABLES SHOWS THE MOST HIGHLY MENTIONED SPECIFIC AREAS, I MENTIONED MORE THAN 200 ALL TOGETHER. SHOW THE -- THE VALUE OF THE RESEARCH COMMUNITY GOING DOWN, THE CURSOR HERE, I'LL GET IT. OOPS. MAYBE I WON'T. GOING DOWN THE FIRST COLUMN, THE VALUE TO THE RESEARCHER, THERE WAS A PUBMED CENTRAL NCBI, CLINICALTRIALS.GOV AND MESH AND SO ON. THE BRIGHT COLORS REPRESENT SERVICES THAT WERE MENTIONED ACROSS THREE OR FOUR DIFFERENT GROUPS. SO IF YOU START WITH PUBMED, IT WAS MENTIONED IN ALL FOUR GROUPS AND IT'S REFLECTED IN THE ORANGE BROWN. AS WAS PUBMED CENTRAL. AND CLINICALTRIALS.GOV. THOSE MENTIONED BY THREE OF THE GROUPS, INCLUDE HISTORY OF MEDICINE, PATIENTS AND THE LIBRARIANS. MEDLINE PLUS, AND DOC LINE. AND THEN THE OTHERS IN THE LIGHT BLUE GRAY WASHED OUT WERE UNIQUE TO EACH OF THOSE COMMUNITIES AND NOT THAT SURPISING FOR THE RESEARCH COMMUNITY OF COURSE NCBI CAME IN, WITH TWO VOCABULARY SERVICES MESH LMS, COCKPIT FOR HEALTH PROFESSIONALS, THREE EMERGENCY TOXICOLOGY RESOURCES CAME IN, WISER TOX SMITH AND RECOMMEND PLUS OPEN ACCESS. FOR PATIENTS PULLED IN THREE CONSUMER HEALTH RESOURCES, K-12 RESOURCES, HOUSEHOLD PRODUCTS AND PUBMED HEALTH. DATABASE TRAINING, WHICH REFLECTS THE OUTREACH THAT WE DO TO THAT COMMUNITY. THEN FOR LIBRARIAN COLUMN, THESE ARE ATTRIBUTED TO THE LIBRARIAN COMPONENT OF THAT COLUMN PULLED IN MESH DATABASE TRAINING RESOURCE TRAINING, AND LIBRARY ASSOCIATES PROGRAM. HELLO BACK THERE. THIS ONLY REPRESENTS LESS THAN 20 ELEMENTS AS THEY REFERRED TO THEM OF NLM. IF WE HAD MORE THAN 200 AREAS REPRESENTED. SO THIS GROUP FOR INTERNAL ANALYSIS, NOT BY VOLUME BUT PROGRAMMATIC AREAS AND THIS REFLECTS THE RANGE OF THE TOPICS OF THE COMMENTS. SO BIBLIOGRAPHIC INCLUDES PUBMED AND PUBMED CENTRAL BUT ALSO THINGS LIKE OKAY ACCESS, PUBLIC ACCESS, FREE FOLD TEXT, JOURNAL SELECTION FOR MED LINE, TOPICS RELATED TO ACCESS TO BIB BLOW OWE GRAPHIC RESOURCES. BIB BLOW OWE SOME MAY SEEM OF COURSE BUT I MENTION TO REALLY GIVE YOU AN IDEA OF THE RANGE OF THINGS DISCUSSED. THAT SAID, CLINICALTRIALS.GOV, DATA SCIENCE AND STANDARDS ARE STRAIGHT FORWARD. THE EHR AND LAC RND, ELECTRONIC HEALTH RECORDS, IN WHICH LISTER HILL IS HIGHLY INVOLVED AS WELL AS OTHER INTRAMURAL INFORMATICS R&D PROJECTS WITHIN THE LISTER HILL VENTURE. INFORMATICS RESEARCH AND INFORMATICS TRAINING REFERS PRIMARILY TO THE EXTRAMURAL PROGRAM. BUT ALSO INCLUDES THE BIOMEDICAL INFORMATION TRAINING COURSE IN GEORGIA FORMALLY WITH HELL. COLLECTIONS REFERS TO BOTH PRINT AND DIGITAL COLLECTIONS AS WELL AS ACCESS TO THE COLLECTIONS INCLUDING THINGS LIKE THE NLM CATALOG, STOCK LINE AND LIBRARY O LIBRARIAN TRAINING REFERS TO DATA BAIN TRAINING WORK SHOT SHOPS AS WOMEN AS DEVELOPMENTAL TRAINING PROGRAMS HISTORY OF MEDICINE, IT UP COLLUDED -- INCLUDES THE EXHIBITION PROGRAM, COMMENTS RELATED TO THAT. AS WELL AS HISTORICAL PRINT AND DIGITAL COLLECTION. CONSUMER HEALTH NOT ONLY COVERS MEDLINE PLUS, GENETIC HOME REFERENCE, PUBMED HEALTH AND OTHER CONSUMER HEALTH RESOURCES BUT ALSO OUR SPECIAL POPULATIONS WEBSITE AND OUTREACH TO HELP DISPARITY AND UNDERSERVED POPULATIONS. THE ENVIRONMENTAL PIECE IS FAIRLY STRAIGHT FORWARD, REPRESENTING THE WIDE RANGE OF SERVICES FROM SPECIALIZED INFORMATION SERVICES. BUT IT ALSO INCLUDES REUNIFICATION PROGRAMS THAT ARE SUPPORTED AND OTHER DISASTER RELATED PROGRAMS THAT ARE SUPPORTED IN THE LISTER HILL CENTER. HEALTH SERVICES RESEARCH ENCOMPASSES PUBLIC HEALTH, BOTH OF THOSE AREAS ARE COMBINED IN THE EHSR OFFICE. THE API CATEGORY CAPTURE IT IS RESOURCES REPROVIDE TO HELP OTHERS USE THEIR LOCAL PROGRAMS TO ACCESS OUR DATA, OR LINK TO RESOURCES WE PROVIDE AND NLM OVERALL IS REALLY THE NLM IS GREAT CATEGORY. THIS REFERS TO THE MANY COMMENTS THAT OFFER GENERAL PRAISE FOR NLM AND ITS RANGE OF SERVICES AND NLM DATABASES. SO IN ADDITION TO CHARACTERIZING THE REASON THESE SERVICES WERE OF VALUE TO THEM, THE RESPONDENTS ALSO OFFERED MANY SUGGESTIONS THAT GENERALLY FALL INTO THESE CATEGORIES. EXPAND, EXPAND OR CONTINUE TO SUPPORT, ADVERTISE, THERE WAS AS YOU MIGHT -- YOU MAY NOT BE SURPRISED TO HEAR THAT THERE WERE A LOT OF COMMENTS IS A SAID OUR RESOURCES WERE -- THE PEOPLE WERE INSUFFICIENTLY AWARE OF THE RESOURCES THAT WE HAVE AND WE SHARE THEM MORE BROADLY. TO IMPROVE AT OR RESTORE FEATURE, AUGMENT CONTENT, CREATE AN APP, COLLABORATE WITH OTHER GROUPS, AND INCREASE FUNDING SO THESE WERE THE KIND OF GENERAL CATEGORIES OF SUGGESTIONS THAT WE RECEIVED. BUT NOW TURNING TO QUESTION 5 WHICH ASKS HOW NLM COULD BE BETTER POSITIONED TO ADDRESS CHALLENGES IN AREAS SUCH AS BIOMED -- ENVIRONMENTAL INFORMATICS. BIG DATA, EHR, DIGITAL PUBLICATION. I WOULD LIKE TO HIGHLIGHT SOME OF THE MAIN THEMES SO IN ADDITION TO THOSE THAT WERE EXPLICITLY ASKED IN THE QUESTION, COMMENTS WERE ALSO OFFERED RELATED TO CLINICAL GENOMICS, CLINICAL DATA ANALYSIS, OPEN ACCESS AND EMERGENCY RESPONSE. SO THIS SECTION WAS REALLY FILLED WITH A VERY RICH SET OF COMMENTS COVERING MANY TOPICS AND THESE WERE JUST SOME OF THE COMMENTS, COMMENTS EXPRESSED THAT I THOUGHT WOULD BE APPROPRIATE TO BRING TO THE ATTENTION OF THE BOARD IN THIS PARTICULAR FORUM. WHILE THERE WAS GENERAL AGREEMENT SOMETHING NEEDS TO BE DONE IN THESE AREAS, THERE WAS DECIDEDLY NOT OVERWHELMING AGREEMENT ABOUT WHO DOES WHAT, THE BEST WAY TO GO ABOUT IT. AND APPROPRIATE ROLE FOR THE DIFFERENT PLAYERS. SO I'LL HIGHLIGHT SOME OF THOSE AREAS OF PENSION AS I GO ALONG. THAT'S PRECISELY WHAT MAKES ALL THESE AREAS INTERESTING IN TERMS OF CONSIDERING THE VARIOUS OPTIONS FOR ACTION AND PROGRESS AND PLANNING, AS WE GO FORWARD IN THIS JUNCTURE FOR THE NLM. NOW, THERE'S SOME OF THESE BULLETS REPRESENT SOME DENSE CONCEPTS UNDERNEATH THEM SO I MAY RELY A LITTLE BIT MORE HEAVILY ON MY NOTES HERE IN THIS PARTICULAR SECTION. SO COMMENTS ON BIOMEDICAL INFORMATICS RELATED TO SUPPORT FOR RESEARCH AS WELT AS TRAINING AND WORK FORCE DEVELOPMENT. THERE WAS STRONG STRONG RECOMMENDATIONS TO INCREASE FUNDING WITH AT LEAST ONE RESPONDENT USING THE PHRASE GROSSLY UNDERFUNDED, TO CHARACTERIZE THE CURRENT SITUATION AND THAT PHRASE -- THE TONE OF THAT PHRASE WAS ECHOED THROUGHOUT. EMPHASIS WAS PLACED ON NEED FOR RESEARCH PORTFOLIO THAT IS BOTH BROAD AND DEEP, BOTH FUNDAMENTAL AND APPLIED, AND BALANCED TO INCLUDE CLINICAL INFORMATICS BIOINFORMATICS AND PUBLIC HEALTH INFORMATICS. SPECIAL IMPORTANCE OF FUNDAMENTAL RESEARCH AND INFORMATICS AS NLM PRIORITY WAS MENTIONED NOTING THAT FUNDAMENTAL RESEARCH IS IN CRITICAL NEED OF CONTINUOUS STRENGTH AND SUPPORT. AND MAYBE UNDERAPPRECIATED IN OTHER VENUES. THE NEED FOR GREATLY EXPANDED TRAINING WAS MADE ABUNDANTLY CLEAR NOTING THE TOUGH CUTS IN THIS AREA IN RECENT YEARS. SOME SAID CONCENTRATE ESSENTIALLY ALL THE EXTRAMURAL FUNDING IN NLM CREATING A LARGER FOOTPRINT AND RECOGNIZING NLM AS THE CENTRAL BROADLY RECOGNIZED HOME FOR INFORMATICS AT NIH. OTHERS SAID DISTRIBUTING FUNDS -- FUNDING BROADLY THROUGH THE ICs AT NIH AND CHANGING CURRENT ATTITUDES TOWARD FUNDING INFORMATICS IN THEIR RESEARCH PORTFOLIOS WOULD BE MORE EFFECTIVE AND HEIGHTENING THE PROFILE OF INFORMATICS AS A CENTRAL RESEARCH AREA IN TODAY'S SPECIFIC ARENA. MOVING TO BIG DATA, RECOMMENDATIONS WERE MADE AND ISSUES RAISED REGARDING THE BUILDING AND HOSTING OF REPOSITORIES INTEGRATIVE SYSTEMS TO FACILITATE DATA, CREATING THE FINDING TOOLS, AND DEVELOPING CONVENTIONS AND CHANGING CULTURE AROUND DATA CITATION AND SOFTWARE CITATION. THERE WERE ALSO COMMENTS ABOUT THE NEED TO ENFORCE DATA TRANSPARENCY, IMPLEMENT STANDARDS, DEVELOP PROTECTIONS AGAINST FRAUD AND REQUIRE DATA MANAGEMENT PLANS. THE CONCEPT OF PARTNER IS MENTIONED THREE TIMES ON THIS SIDE OF THE SLIDE. THIS ISEN AREA I REFERRED TO AS I'M SURE YOU ARE AWARE. SOME RESPONDENTS AT NIH SHOULD DO THESE THINGS. OTHERS NOTED EXTERNAL GROUPS HEAVILY ENGAGED IN A PRODUCTIVE WAY IN THESE AREAS. AND URGE THAT NIH NLM RECOGNIZE THOSE ACTIVITIES AND PROCEED IN THE STYLE OF PARTNERSHIP AND COLLABORATION. ONE OF THE COMMENTS RELATED TO AREA OF CLINICAL GENOMICS ACTIVITY AT NIH, AND RAISE THE ISSUE OF WHERE NLM INFORMATION SYSTEM SHOULD BE ENGAGED FOR CLINICAL USE. CLINICAL INFRASTRUCTURE AND CENTRALIZED MANAGEMENT ANALYZED SHOULD NIH AND NLM MOVE IN THIS DIRECTION. IN THE MORE GENERAL AREA OF CLINICAL DATA ANALYSIS THERE WERE RECOMMENDATIONS REGARDING ELEMENTS TO SUPPORT THE SECONDARY ANALYSIS OF CLINICAL DATA, THESE INCLUDED CREATING A CENTRAL REPOSITORY OF PRIMARY DATA THAT COULD BE ANALYZED. AND DEVELOPING ONLINE NAVIGATION TOOLS AND VIEWERS AKIN TO THOSE DEVELOPED BY NLM FOR BROWSING GENOME DATA. THE NEED FOR CREATE COMMON REPORTING STANDARDS FOR VARIOUS HEALTH RELATED EXPOSURES WAS ALSO NOTE AS IMPORTANT PREREQUISITE FOR META ANALYSIS. SO EHR GETS BOTH COLUMNS ON THIS SLIDE. IN THE AREA OF ELECTRONIC HEALTH RECORDS THERE WERE MANY COMMENTS AND RECOMMENDATIONS ABOUT VOCABULARY AND OTHER DATA STANDARDS, PRAISING NLM LEADERSHIP IN THIS AREA, BUT NOTING THAT CONTINUING GREAT NEED ALSO IN THIS AREA. THE WORDS COLLABORATE AND PARTNER FIGURE STRONGLY IN THIS ARENA WITH SOME SAYING NLM SHOULD TAKE THE LEAD DO IT ALL, OTHERS SAY CONTINUE WITH SNOW MED AND OTHER CURRENT ACTIVITIES BY BACK OFF AND RECOGNIZE THE ROLE OF OTHERS IN DEVELOPING NEW STANDARDS AND VALUES THAT INCLUDED PRIVATE SECTOR. THREE SPECIFIC AREAS IN NEED OF STANDARDS WERE NOTED. DRUG ALLERGY, ADVERSE DRUG EFFECTS AND BIOTECHNOLOGY DERIVED PHARMACEUTICALS. THEN THE RIGHT HAND COLUMN, COMMENTS REFLECTING MORE OF THE SYSTEM DEVELOPMENT SIDE ARE NOTED. THERE WERE SUGGESTIONS TO CREATE ENTIRE TEST BED SYSTEM WHICH COULD ALSO BE USED FOR VARIOUS DEVELOPERS TO CREATE APPLICATIONS TO TEST AND DEMONSTRATE THE VALUE OF THE DIFFERENT PLUG INS AT POINT OF CARE. THERE WERE SUGGESTIONS TO CREATE DATABASES OF PATIENT DATA FROM E HEALTHCARESRs, LINKED FOR RESEARCH AND ANALYSIS PATTERNS. PARTNERS PLUM METED ON THESE COMMENTS WITH PUSH ON ONE HAND FOR NLM TO LEAD AND PUSH FOR OUTSIDE GROUPS ON THE OTHER. RELATE ODD TO THE DEVELOPMENT OF EHR SYSTEMS WAS ALSO THE OBSERVATION A BRIDGE TO KNOWLEDGE BORROWED FROM THE COMMENTS YESTERDAY, TRAY PHRASE, BRIDGE TO KNOWLEDGE WAS ALSO IMPORTANT FOR CLINICAL INFORMATICS. IN THIS REGARD NEED TO DEVELOP AND FUND CLINICAL DECISION SUPPORT SYSTEMS THAT WOULD BE INTEGRATED WITH THE EHRs DESCRIBED. THE AREA OF DIGITAL PUBLICATIONS OVERWHELMINGLY TO RETAIN PRINT COLLECTIONS AS NATIONAL LIBRARY. THAT SAID THERE WERE STRONG SUPPORT TO DIGITIZE AS MUCH AS POSSIBLE MODERN HISTORICAL COLLECTIONS, ALONG WITH RECOMMENDATIONS, TO CAPTURE THE CURRENT WEB CONTENT THROUGH WEB ARCHIVING ACTIVITIES AND THE CURRENT EFFORTS NLM IN BOTH AREA S HIGHLY RECOGNIZED. IN THE HISTORY OF MEDICINE AREA, THERE WERE RECOMMENDATIONS TO UNDERTAKE LARGE SCALE ACQUISITIONS PROJECTS FOR PRE-1920 MEDICAL RECORDS AND ALSO TO UNDERTAKE SCANNING PROJECTS 3-D SCANNING PROJECTS FOR HUMAN ANATOMICAL COLLECTION. REGARDING OPEN ACCESS THE SUPPORT FOR CONTINUED EXPANDED OPEN ACCESS WITH FREE PULL TEXT ARTICLES AVAILABLE TO ALL STRONG THROUGHOUT. THERE WERE REQUEST TO DEVELOP AFFILIATION ROOTS TO FACILITATE THIS PARTICULARLY FOR PUBLIC HEALTH WORKERS, SCHOOL NURSES AND OTHER GOVERNMENT HEALTHCARE PROVIDERS. AS WELL AS UNAFFILIATED HEALTH PROFESSIONALS. THERE WERE ALSO RECOMMENDATIONS TO MAKE LINK TO FULL TEXT ARTICLES MORE SEAMLESS. ON THE PROVIDER SIDE THE TEXT TENSION OF WHO DOES WHAT CAME UP AGAIN AS YOU EXPECT DETERMINING RELATIVE ROLES IN THE PUBLISHING AND ARTICLE ACCESS PROCESS PUBLISHERS PRAISED -- RAISE ISSUES RELATED TO COPYRIGHT, ACCESS TO USE MEASURES, POLICIES ABOUT LINKING TO THE FULL TEXT ARTICLE OF RECORD, DEVELOPING POST PUBLICATION SERVICES, AND TAKING A COLLABORATIVE APPROACH TO DEVELOPING DATA SHARING POLICIES. OTHER ORGANIZATIONS SUCH AS THOSE INVOLVED IN UNIQUE IDs FOR AUTHORS AN SYSTEMS FOR MEASURING AND REPORTING USE STATISTICS, ALSO RAISED ISSUES RELATED TO ASSESSING APPROPRIATE ROLES AND WORKING IN PARTNERSHIP. FINALLY ONE OF THE THEMES IN THE OTHER CATEGORY WAS EMERGENCY RESPONSE. TOXICOLOGY EMERGENCY RESPONSE AND DEARS INFORMATION SERVICE WAS STRONG THROUGHOUT THE REQUEST FOR INFORMATION AND IN THE AREA ADDRESSING GROWING CHALLENGES HOWEVER THERE WERE THREE MAIN IDEAS. FIRST SOCIAL MEDIA IS AN IMPORTANT COMMUNICATION CHALLENGE IN TYPES OF DISASTER DESCRIBED AND ACKNOWLEDGED BUT IT WAS NOTED NO FORMAL SYSTEMS SET UP TO IMPLEMENT THAT. THEREFORE A RECOMMENDATION THAT NLM WORK WITH THE COMMUNITY TO DEVELOP A FORMAL MODEL FOR NATIONAL USE OF SOCIAL MEDIA. WITH ROLES AN RESPONSIBILITIES ASSIGNED TO DIFFERENT GROUPS AND VERY DETAILED PLAN TO IMPLEMENT SUCH SERVICE WAS OUTLINED IN THE COMMENT. NEXT LACK OF SUFFICIENT SUPPORT TO RESPOND SAFETY INCIDENCE IN RURAL AREAS WAS ALSO DESCRIBED AND A SUGGESTION WAS MADE TO DEVELOP ONLINE CERTIFICATION PROGRAM. TO FACILITATE NEEDED,PER TEASE IN RURAL AREAS. FINALLY THE NEED TO DEVELOP INFECTIOUS DISEASE INFORMATION SERVICE SIMILAR TO THOSE RADIATION AND CHEMICAL EVENTS REM AND CHEM WAS DESCRIBED. I MUST SAY COMMENTS FROM THIS COMMUNITY THROUGHOUT WERE SOME OF THE MOST MOVING AND COMPELLING. TO HEAR FROM FIRST RESPONDERS ABOUT THE CRITICAL ROLE OUR RESOURCES PLAY EVERY DAY IN THE WORK THEY DO WAS A PARTICULAR PRIVILEGE. SO IN SUMMARY, THE RESPONSES TO THIS RFI DEFINITELY DEMONSTRATED THAT WE HAVE THE HIGHLY ENGAGED USER COMMUNITY, PEOPLE GREATLY INVEST MISDEMEANOR THE SERVICES WE PROVIDE AND THE RESEARCH WE SUPPORT. THEY OFFER THOUGHTFUL INPUT WITH RELEVANCE ACROSS NLM. THE PROCESS TEAM THROUGHOUT THE LIBRARY SPECIFIC FEEDBACK AND SUGGESTION AS THEY MAKE MANAGEMENT DECISIONS. SUGGESTIONS FOR FUTURE DIRECTIONS WILL CONTRIBUTE GREATLY IN DEVELOPING PROGRAM PRIORITIES. I THINK I MAY HAVE FAILED TO MENTION AT THE OUTSET THAT AS DR. COLLINS MENTIONED YESTERDAY THESE RESPONSES WERE CONSIDERED VERY SERIOUSLY BY HIS WORKING GROUP AND FIGURED STRONGLY IN THE REPORT AND RECOMMENDATIONS THAT THEY PROVIDED AND THAT HE ACCEPTED. S FROM WE ALSO LOOK FORWARDED TO CONTINUED DISCUSSION WITH BOARD ON THESE ISSUES AS NLM STRATEGIC PLANNING GOES FORWARD. SO LIKE TO CLOSE BY THANKING THE BOARD FOR EFFORTS IN COMMUNICATING WITH THE VARIOUS NLM USER COMMUNITIES AND THE RFICS ENCOURAGING WIDESPREAD PARTICIPATION WHICH WE CERTAINLY ACHIEVED. AND I WOULD ALSO LIKE TO ACKNOWLEDGE AND THANK OTHER NLM STAFF ASSISTING IN ANALYSIS OF THIS DATA. DANA (INDISCERNIBLE) WORKS WITH ME IN OFFICE OF HEALTH INFORMATION PROGRAMS DEVELOPMENT, MARIA COLLINS AND CASTLE DUNN CONTRIBUTE THROUGH LIBRARY OPERATIONS. MARY ANN LEONARD THROUGH OFFICE OF COMMUNICATION AND I WAS FORTUNATE TO ALREADY ENLIST THE ASSISTANCE OF OUR THREE NEW ASSOCIATE FELLOWS. ALWAYS A NICE PLUS. SO I LOOK FORWARD TO HEARING FROM YOU NOW IN THE DISCUSSION SESSION AND THANK YOU FOR YOUR ATTENTION. [APPLAUSE] >> I HAVE A QUESTION. IS THIS POINT TO BE SHARED -- ANY OF THIS INFORMATION SHARED WITH RESPONDENTS OR HAS IT ALREADY BEEN SHARED WITH THE RESPONDENTS TO THE RFI? >> IT HAS NOT BEEN SHARED WITH THE RESPONDENTS AND I'LL LET -- >> WE HAVE TO TWO BACK, I THINK WE CHECKED ON THIS, WHEN PEOPLE SUBMIT THE RFI ACTUALLY SAID HOW NIH MIGHT USE THE INFORMATION, SOME MIGHT BE PRIVATE OR SOME MADE PUBLIC. YOU MAKE A GOOD SUGGESTION. (OFF MIC) >> THEY DON'T HEAR ANYTHING ABOUT WHAT -- I THINK THAT -- I DON'T THINK THERE'S ANY REASON NOT TO PUT UP -- AS YOU CAN SEE IT REMINDS ME SOME OF THE BIG ISSUES THAT SOME OF THE THINGS THEY -- ARE SAYING WHY DON'T YOU DO X OR Y AT THE END. I THINK THIS IS GOING TO FEED INTO EXECUTION PLANNING ONE RECORD APPOINTED. I WAS STRUCK BY THE PEOPLE THAT ALL THESE CLINICAL GENOMICS ARE TERRIFIC. BUT YOU NEED TO HARDEN THE SYSTEM SO THAT YOUR RESOURCES AS YOU PROVIDE THEM NOW CAN BE ESSENTIALLY INTEGRATED INTO OUR CLINICAL SYSTEM HERE AT THE HOSPITAL. IT'S AN INTERESTING CONCEPT IT'S DEFINITELY DEBATABLE WHERE THAT MAKES SENSE TO DO. BUT I THINK IT'S ONE OF THE NUMBER OF BIGGER QUESTIONS ABOUT WHETHER THAT'S A REASONABLE PLACE FOR NLM TO GO OR NOT OPPOSED TO JUST DATA ARE AVAILABLE AND PEOPLE ARE GOING TO BE ABLE TO POTENTIALLY ADD VALUE-ADDED SERVICES AND NOT NECESSARILY FEEL VERY STRONG REQUIREMENT TO KEEP OUR SYSTEMS UP 24/7. IF PMI TAKES OFF, -- >> WHEN PMI TAKES OFF, THE NOTION THAT EVERY HOSPITAL OR CLINICAL OPERATION IN THE WORLD IS EXPECTING TO VALIDATE DATA OR DO SOMETHING DIRECTLY OFF THE NLM SYSTEM, THERE IS AN ISSUE PROBLEM OR DELAY WHATEVER THEY HAVE PATIENT THERE IS IN FRONT OF THEM. THAT DOESN'T NECESSARILY STRIKE ME AS A USEFUL PLACE TO GO. >> SO YOU (INAUDIBLE) CIRCULATE TO BETSY'S -- IT SORT OF THE RELATES TO BETISES'S COMMENT ABOUT DECISION SUPPORT AND I WONDER IF THERE'S ANY EXPLICIT COMMENTS ABOUT KNOWLEDGE BASES ESPECIALLY COMPUTABLE KNOWLEDGE BASES AS WE GO MORE AND MORE KNOWLEDGE ARTIFACTS AND BEST PRACTICE RECOMMENDATIONS, THEY'RE NEVER HARDEN BECAUSE THEY KEEP CHANGING. POTENTIAL RECOMMENDATION. THERE REALLY -- WE HAVE TO SCATTER PLACES WHERE THESE RESIDE NOW, THERE'S NO AUTHORITATIVE PEER REVIEW PROCESS MANAGED BY SOME IDENTITY PUBLIC PRIVATE OR ACTUALLY HOUSE THESE IT OFTEN COMES UP WHERE SHOULD THESE LIVE? OUR CLINICAL -- -- PCORI IS THIS A ROLE FOR NLM TO -- I WANT PEOPLE TO COMMENT. >> IT WAS REALLY -- I MADE A NOTE OF IT AND I WILL SEARCH TO TEXT KEY WORDS AROUND THAT TO TEASE OUT BUT I DO NOT RECALL SEEING THAT CONCEPT ISSUE, IT MIGHT BE BURIED IN A PHRASE, IN A SERIES OF MENTIONS. >> PEOPLE CERTAINLY MENTIONED IT DIRECTLY TO ME. >> YES. >> OTHERS ON THE STAFF MAY HAVE HEARD THE SAME THING. >> (INAUDIBLE) BD 2K, HOW IS IT MANAGED AND HOW IS IT CURATED, THE TOOLS AND METHODS FOR THAT, AND MAYBE THE REPOSITORIES UNDER SOME -- I WOULD SUBMIT THAT IS (INAUDIBLE). IT'S YET TO BE -- >> IT'S INTERESTING BECAUSE SOME OF THE -- SOME OF THE COMMENTS HERE -- SOME OF THE THINGS (INAUDIBLE) WHATEVER, I THINK PIECES OF WHAT THEY'RE TALKING ABOUT HERE ARE LIKE THE OUTCOMES OF THE OTHER STATISTICAL METHODS INITIATIVE IN TERMS OF CERTAIN TYPE OF DATA AVAILABLE TO RESEARCHERS AND WE'LL SEE THAT UNFOLDED. BUT -- >> WE HAD -- THE CHALLENGE OF HEALTH INFORMATION EXCHANGE BUSINESS MODEL STILL SURVIVE. SOME DO OKAY IN GENERAL BUT THE HIU, THE IMPORTANT NATIONAL RESOURCE, WHICH IS WHAT YOU WANT TO DO AND WE DON'T HAVE -- HIV CONSUMER GENERATED PATIENT ACCESS DATA. I THINK BOTH THE DEVELOPMENT OF PMI AND DEVELOPMENT OF THE NLM STRATEGIC LONG RANGE PLAN WILL HAVE A LOT OF INTERESTING CHALLENGES. AND IN THE MARKET THAT CAME UP WITH THE SENATE TESTIMONY YESTERDAY ABOUT PATIENT ACCESS TO DATA, EVERY INDIVIDUAL SOFTWARE SERVICE PROVIDER, MANY MEDICAL PROVIDERS ARE STARTING TO TREAT PATIENT DATA LIKE THEIR IP AND THEY WANT TO MONETIZE IN DIFFERENT WAYS, BECOMES MORE DIFFICULT FOR PEOPLE TO GET ACCESS TO THEIR OWN DATA. WHAT THE DATA AND KNOWLEDGE BASES, I THINK IT'S PROVOCATIVE AND CHALLENGING, IT'S DIFFERENT INFRASTRUCTURE SUPPORT TO DO IT BUT SOMEBODY HAS TO COUNTER ACT THE SORT OF MONETIZATION PUSH WITH THE BUSINESS PUSH. >> I THINK THAT TO THAT POINT, WITH OUR UNDERSERVED, PARTICULARLY WITH PUBMED, MEDLINE PLUS AND SOME OF OUR SERVICES THERE HAS ALWAYS BEEN FOR MANY, MANY YEARS OBVIOUSLY THERE HAVE BEEN THIS ISSUE OF I THINK WE HAVE BEATEN IT TO DEATH BUT INITIALLY AFTER THE PIONEERING WORK OF BOTH MED LINE TO GIN WITH TO THE MARRY ONLINE SERGING WITH COMMUNICATIONS WHICH IS THE EXPERIMENT CARRIED OUT BY THE GENERAL CENTER WHICH RESULTED IN MED LINE, AND SUDDENLY EVERYBODY REALIZES REALLY A MARKET FOR THIS AND IT COULD ACTUALLY BE DONE. WITHIN FIVE, SIX, SEVEN YEARS OF US DOING IT THAT'S WHEN WE BECAME VERY UNFAIR FOR COMPETITION IN THE PRIVATE SECTOR CLEARLY MED LINE COULD BE A HUGE -- IF WE WOULD JUST PRODUCE THE DATA AND GET IT TO US SO THAT WE CAN MAKE IT AVAILABLE. AT THAT TIME CHARGING BECAUSE WE HAD TO PAY THE COMMERCIAL TELECOM BILL WE HAD A ALGORITHM BASED ON AMOUNT OF USE THAT KEPT THE PRICE FLOW LOW. WHAT WE PROVE WHEN DAWN WAS ABLE TO ORCHESTRATE THROUGH THE HELP OF FRIENDS DROPPING THE CHARGE BECAUSE OF EXPANSION OF INTERNET WHICH MAY OF COURSE TECHNICALLY POSSIBLE AND WE ENLISTED FRIENDS AND PEOPLE (INAUDIBLE) HARKEN WHATEVER TO GIVE US THE COVER TO DO IT. WE OF COURSE LEARN MONTH DECEMBER CHARGE WAS KEEPING PEOPLE WHERE FROM USING IT BECAUSE -- BUT IN EFFECT WHAT WE WERE WAS WE WERE KEEPING THE PRICE REASONABLE ACROSS THE BOARD. EVEN IF YOU CAN PROVIDE A BETTER SERVICE, I HAD TO BE A SERVICE -- YOU COULDN'T BE SO FAR OUT OF TESTIFY PRICE THAT WE WERE PROVIDING OR YOU HAD TO HAVE A VERY SIGNIFICANT VALUE FOR SOMETHING. YOU COULD HAVE A SO-SO NOT VERY GOOD SERVICE AND CHARGE A LOT BUT PEOPLE WOULD TURN AND SAY I REALLY LOVE THE DIALOGUE BETTER THAN -- BUT ON THE OTHER HAND, THE PRICE DIFFERENTIAL WE CAN'T AFFORD IT, THIS ONE IS GOOD ENOUGH. SO I FEEL LIKE WE ARE THERE TO PROVIDE SERVICES AND DATA IN WAYS AND THE FACT WE'RE BOTH RETAIL AND WHOLESALE, HAS A TOTAL IMPACT ON ACCESS TO HEALTH INFORMATION AND SCIENTIFIC INFORMATION. WE DON'T HAVE TO HAVE EVERY -- WE DON'T HAVE TO HAVE EVERY NICHE MARKET, WE DON'T HAVE TO HAVE WHATEVER BUT WE HAVE TO HAVE THE ABILITY FOR INDIVIDUALS TO GET INFORMATION T IF COSTS REALLY MATTER TO THEM OR THEY'RE IN AN ENVIRONMENT WHERE PARTICULAR NICHE PRODUCT IS NOT AVAILABLE TO THEM. TRZ SO WE -- I -- THAT'S HOW WE HAVE PLAYED IT UP TO NOW. AND IT'S BEEN A GOOD WAY TO PLAY IT BUT THESE ISSUES ARE SERIOUS ISSUES AN DESERVE A LOT OF DISCUSSIONS AN STRATEGIC PLANNING FORWARD THAT MAKES TIME FOR A CHANGE. >> THAT'S WHERE WE'RE GOING TO HAVE TO REALLY AS A BOARD AND OFFICE OF PLANNING, -- ABOUT HALF THESE RECOMMENDATIONS CIRCULATE AROUND NEW RECORDS OF KNOWLEDGE. AND YOUR ILLUSION TO MED LINE TAKES US BACK TO IN MY MIND IMMEDIATELY SUPPORT. AND YOU HAVE A NEW RECORD OF KNOWLEDGE IN THE PEER REVIEW PUBLICATIONS OF JOURNALS AND A WAY OF GETTING NOW. WE WENT WHOLESALE WELL BEFORE WE WENT RETAIL. AND WE GOT MOST OF THE BUGS OUT AT THE WHOLESALE LEVEL. SO WHEN WE WANT TO GO RETAIL, WE COULD IN FACT DO IT AT A REASONABLE LEVEL AND IF YOU HAD CURRENT CONTENTS OR SOMETHING ELSE THAT HAD ADDED VALUE YOU COULD SELL THAT BUT AS YOU SAY IT KEPT THE PRICE IN A MANAGEABLE RANGE. SO WE NEED TO LOOK AT THESE NEW RECORDS OF KNOWLEDGE THAT SAME WAY, WHO NEEDS IT, WHAT DO THEY NEED HOW IS IT LIKELY USED IF WE ARE IN COMPETITION WITH REAL BUSINESS, THEN THERE'S A LAW THAT SAYS THE GOVERNMENT OUGHT NOT DO THAT. WE NEED TO BE DAREFUL ABOUT THOSE THINGS. >> I AGREE WITH YOU ENTIRELY, DALE. I THINK THAT'S VERY GOOD POINT. THE REASON WHY WE HAVE WON THE ARGUMENT PERIODIC AS THEY ARE, IS IN GENERAL IT'S VERY EASY FOR US TO PROVE THAT IN TERMS OF DEVELOPMENT OF BROADLY AVAILABLE SERVICES IN THE AREA, WE WERE THERE FIRST. SO IF WE WERE TO SAY WELL THERE ARE TWO OR TREE VERY GOOD SERVICES IN ALL HOSPITALS IN THE UNITED STATES HAVE THEM, BUT THEY COST A LOT OF MONEY, BUT THE HOSPITALS ON THE OTHER HAND ARE PAYING FOR THEM AND PEOPLE REVIEW THEM NOW BUT HEY MAYBE WE CAN CREATE A BETTER ONE. THAT I THINK IS THE LESS -- BETTER THING -- >> SHOULD BE A NICE SIDE EFFECT BUT NOT -- PEOPLE'S ACCESS TO DATA MARKET IMPACT, THAT'S A GOOD THING BUT -- >> RIGHT EXACTLY. IT'S NOT THAT WE HAVE BEEN SHY ABOUT ABOUT DOING THINGS THAT WE THOUGHT WERE HIGHLY BENEFICIAL IN TERMS OF ACCESSING INFORMATION THAT DIDN'T STRIKE CERTAIN ELEMENTS WHERE THERE WERE CERTAIN ELEMENTS THE INFORMATION ARE ELSEWHERE WE WOULD HAVE BEEN JUST AS -- I THINK WE HAVE ALSO IN MANY CASES, I HAVE BEEN IN A NUMBER OF CONVERSATIONS OVER MY DECADE HERE AT NLM, SOMEBODY CAME UP WITH A GREAT IDEA, NLM SHOULD DO X AND IN FACT THERE WAS A DISCUSSION OF LOOKING AROUND VERY WELL DONE SOMEWHERE ELSE, AS WE WOULD NOT BE PLUGING A HOLE THAT WAS IMPORTANT. >> AS AN ASSAYED. THINGS THE IP -- THE PRIVATE SECTOR GET IP IMPLEMENTATION DATA. WASN'T (INAUDIBLE) GENOME? >> YES. >> SO THAT NOT A PRESS PRECEDENT FOR -- >> PARTLY WHAT IS HAPPENING IN TERMS OF -- JOSH IS HERE WHEN HE TALKS PMI WE HAD A BIG -- PARTICIPANTS GET ACCESS TO THE DATA. WHEN YOU START TO LOOK AT THE VARIETY OF DATA TYPES THAT ARE THERE, THAT THEY'RE NOT GOING OPEN UP API FOR THIS, REALLY GOOD LUCK ACTUALLY GETTING THAT. SO IT'S -- WE'RE GOING TO PLAY THAT OVER AGAIN IN THESE AREAS AN FIGURE OUT -- I MEAN, REALITY IS THE LAW SAYS IF YOU HAVE ACCESS TO THE DATA, GETTING IT IS A WHOLE DIFFERENT STORY BECAUSE OF THE COST, SO FORTH. >> OF COURSE IF YOU HAVE THE RAW DATA YOU -- THE PATIENTS WILL NEED TO GET -- >> IF YOU WANT DIFFERENT ALGORITHMS YOU WANT THE RAW DATA BECAUSE YOU'RE TRYING TO COMPARE STEPS, (INAUDIBLE) MANUFACTURER AND THEY'RE ONLY GOING TO GIVE YOU THE -- HARD TO DO COMPARISONS ACROSS DIFFERENT TYPES OF THINGS FOR RESEARCH PURPOSES. >> ONE OF THE THINGS THAT I FEEL -- I THINK THIS WHOLE NOTION OF ONLY INTERPRETATION GOES BACK TO WHOEVER ORDERED IT AND IN THE CASE OF A LOT OF CLINICAL GENOMIC TESTING AND SO FORTH, WE'RE SAYING IT ONLY GOES BACK TO THE MAJOR MEDICAL CENTER, OR WHATEVER THAT ORDERS THE TEXT, THEY DON'T NECESSARILY GET THEY GET THE INTERPRETATION POTENTIALLY CLINICALLY ACTIONABLE VARIATION BUT THEY DON'T ACTUALLY GET BACK THE DATA THE LAB HAD. >> ISSUE YOU'RE TRYING TO USE THE PATIENT AS THE MEANS TO SORT OF SAY I WANT TO DONATE MY DATA FOR RESEARCH. THEY ACTUALLY NEED ACCESS TO THE RAW DATA. >> THE FACT THE THING THAT'S INTERESTING, THERE WAS A MANY OF YOU MAY KNOW, THERE WAS A REVISIONS JOINTLY TO CLIA AND HIPAA AND BASICALLY SAYS YOUR ENTITLED TO THE DATA AT WHATEVER ANYBODY HAS IDENTIFIABLE FOR YOU, YOU ARE ENTITLED TO IT. SO THE CORRECT INTERPRETATION OF THAT IS IF I WANT MY -- IF YOU DID THE WHOLE GENOME WHATEVER ON ME, AND I WANT THE RAW DATA AS WELL AS INTERPRETATION YOU HAVE TO GIVE IT TO ME. I THINK THAT WHAT WE'LL HAVE IN THIS AREA WE OFTEN HAVE IN OTHER AREAS THE LAW SAYS THIS BUT I HAVE TO GO THROUGH SO MANY DOORS, AND CLIMB OVER WALLS TO GET IT. SO THEN WHAT WE'LL HAVE IS A FEW CASES. AND THEN MAJOR LABS GET OVER THIS. (OFF MIC) >> YOU CAN GIVE IT TO SOMEBODY IF THEY WANTED IT. >> JUST ADD ASSESSMENTS, FOLLOW, VERY USEFUL COMMENTS. WANT TO THANK YOU FOR -- STAFF FOR TAKING THIS UNSTRUCTURED SMALL BIG DATA -- [LAUGHTER] >> AND TO FOLLOW-UP ON THAT, I'M NOT SURE THAT NLM SHOULD BE INVOLVED IN MANAGING ALL THIS BIG DATA. WHAT I WOULD RATHER SEE IS NLM ESTABLISHING STANDARDS HOW BIG DATA STRUCTURED AND HOW IT SHOULD BE EXCHANGED AND LET THE FOLKS OUT THERE, DEAL WIT. MOVING OVER TO THE CLINICAL SIDE, ELECTRONIC HEALTH RECORD PERSONAL INFORMATION ALL OF OUR LABS ALL OF OUR HISTORY, AND CLINICIAN, GIVE US SOMETHING WE CAN PROVIDE TO MEDICAL COMMUNITIES TO FIGURE OUT WAY SEAMLESS EASY TECHNOLOGY, APIs AND ANYTHING ELSE. ALL THE OTHER STUFF I THINK WILL BE -- UNLESS SOMEONE GIVE AS COUPLE OF MILLION DOLLARS. >> I THINK THOSE ARE GOOD POINTS LIKE THEY ALWAYS ARE. I DO THINK THAT THESE ARE THE KINDS OF THINGS WE DISCUSSED. ONE OF THE COMPUTER SCIENCE TELECOMMUNICATIONS COMMITTEE THAT TALKED ABOUT AND I HAVE IT IN MY OFFICE, YOU GET IT. IT WAS ESSENTIALLY WHAT KIND OF COMPUTER SCIENCE RESEARCH SO FORTH NEEDED TO BE FUNDED -- SUPPORTED AND I AM STILL MIGHTILY CONVINCED BY ONE OF THEIR MAIN RECOMMENDATIONS BEFORE THEY CAME UP WHICH WAS ESSENTIALLY THE BIG ISSUE HOW TO SUMMARIZE THE DATA O IT IS PRESENTING A MEANINGFUL PICTURE TO THE PERSON OR TO THE HEALTHCARE PROVIDER. ABOUT IN SOME WAYS THE HEALTH STATUS AND WHAT THE BIG ISSUES ARE AND PERMUTATION POINT OF VIEW SOMEONE MIGHT SAY IF YOU ONLY HAVE TIME TO DO ONE THING FOR YOUR HEALTH, FOR YOU, IT'S THIS ONE. TY STILL FEEL FROM MY PERSPECTIVE IF NLM ACHIEVES WHICH I HOPE WILL, A MUCH LARGER EXTRAMURAL BUDGET, THAT FOCUSING ON SOME OF THOSE ISSUES ABOUT HOW YOU WOULD CONVERT ALL THIS IN TO SOMETHING THAT IS REALLY HELPFUL TO A PATIENT AND A PROVIDER, IN A 15 MINUTE INTERVIEW TO DO BEST FOR THE PATIENT. I THINK THAT THAT CLEARLY IS MORE R&D AND NLM WOULD BE A GOOD PLACE (OVERLAPPING SPEAKERS) >> GOALS, BETSY THE PROVIDER CAN READ AT A LEVEL, WE SEE THAT PROBLEM IN INFORMED CONSENT ALL THE TIME. INFORMED CONSENT FORMS ARE GRADE LEVEL, NONE ARE, MOST HAVE 15 GRADE LEVEL, -- FIFTH GRADE LEVEL. [LAUGHTER] >> (OVERLAPPING SPEAKERS) >> IT'S VERY DIFFERENT. >> I WAS NOT IMPLYING THAT THE SAME SUMMARIZATION WOULD FALL APART, JUST THAT EITHER ONE OF THEM HAS IT NOW. >> I WILL HAVE DR. FRANCIS GIVE ONE LAST COMMENT AND THEN WE NEED TO WRAP IT UP, THIS HAS BEEN A GREAT CONVERSATION. WE HAVE A BREAK. WE WILL ENCOURAGE OTHER (INAUDIBLE). >> I THINK IN ADDITION TO THE PUBLIC GOOD ARGUMENT AND THE ARGUMENT ABOUT WHETHER IT'S GOVERNMENTAL OR NOT, THERE'S DIFFERENT THERAPIES THAT MAY NOT BE AFFECTING NIH QUITE SO MUCH. BUT BELIEVE ME, IT WILL. IT'S AFFECTED US AND THAT'S ASYMMETRY OF OF RISK AND REGULATION APPLYING TO FEDERAL AGENCIES. SO MUCH RISK AVERSION. ONE OF 300,000 EMPLOYEES DOES SOMETHING WRONG IT'S ENTICING THE ENTIRE SYSTEM, BUSINESS HAS LESS OF THAT CONCERN. >> >> (INAUDIBLE) FIVE YEARS AGO WE HAVE -- WE ANNOUNCED A PROBLEM WITH CLEANING ENDOSCOPE AND HEARING HOW INCREDIBLY INCOMPETENT WE ARE, WHEN IT HAPPENS TO VIRGINIA MAYSOR, FDA IT'S YOUR FAULT FOR NOT REGULATING THOSE INSTRUMENTS AND HOLDING THE MANUFACTURERS TO ACCOUNTS SO THAT HAS TO BE FACTORED IN, HOW YOU USE THE INFORMATION AND FRANKLY WE ARE HAM STRUNG. OUR APPROACH NOW IS TO GET DATA OUT IN DAYTIME FORMAT, LET THE PRIVATE SECTOR DEVELOP APPS THINGS LIKE WAITING TIMES. THERE'S TOO MUCH RISK AND SO NOT QUITE THE SAME SITUATION BUT A LITTLE BIT OF CONSIDERATION. >> THANK YOU. [APPLAUSE] >> OUR NEXT PRESENTATION IS BY DR. JOSH DENY AND HE WILL BE TALKING ABOUT PRECISION MEDICINE INITIATIVE ESPECIALLY THE INFORMATICS PIECE. DR. DENNY IS ASSOCIATE PROFESSOR BIOINFORMATICS AND MEDICINE AT VANDERBILT UNIVERSITY AND FORMER NLM FUNDED INFORMATICS RESEARCH FELLOW. DR. DENNY IS PART OF THE NIH FUNDED ELECTRONIC MEDICAL RECORDS AND GENOMICS OR EMERGE NETWORK, AND EMERGED COORDINATING CERTAINTY. THE PHARMACOGENOMICS RESEARCH NETWORK WHICH IS PGRN MORE ACRONYM, AND THE PHARMACOGENETICS OF A VERY LARGE POPULATION, ANOTHER ONE PG NETWORK AT VANDERBILT HE'S ALSO PART OF THE PREDICT WHICH IS PHARMACO GENOMIC RESEARCH FOR ENHANCED DECISION IN CARE AND TREATMENT PROGRAMS WHICH PROGRESSIVE -- PROSPECTIVELY GENOME TYPE -- GENOTYPES PATIENTS TO TAYLOR DRUG RESPONSE. GIVEN THIS RELEVANT EXPERIENCE IT'S NOT SURPRISING THAT HE WAS CHOSEN TO BE A MEMBER OF THE PRECISION MEDICINE INITIATIVE WORKING GROUP OF THE ADVISORY COMMITTEE TO THE NIH DIRECTOR. FOR THE PAST SEVERAL MONTHS DR. DENNY HELD A PART TIME TEMPORARY ASSIGNMENT AT THE NIH ASSISTING WITH THE PREPARATION OF THE WORKING GROUP REPORT WHICH WILL BE PRESENTED AND DISCUSSED BY THE FULL COMMITTEE IN A TELECONFERENCE THIS AFTERNOON. HE WILL BE SPEAKING TO US THIS MORNING ABOUT THE INFORMATICS FOR IMPLICATIONS FOR PRECISION MEDICINE. DR. DENNY. >> THANK YOU VERY MUCH, IT'S PLEASURE TO BE HERE. THE -- I GOT -- YOU READ OFF ACRONYMS. ONE OF THE ROLES OF INFORMATICS IS TO HAVE LOTS OF ACRONYMS, SO -- >> MAKING WORK FOR THE PEOPLE WHO LEARNED HOW TO DETECT THOSE IN FULL TEXT. >> EXACTLY. SO THAT'S WHERE I GOT STARTED. I GET TO EAT MY OWN CHALLENGES. IN FACT AT MY CLINICAL NOTES LOOKING FOR HYPERTENSION I NEVER ONCE SPELLED THE WHOLE WORD OUT IN MY NOTES SO I WRITE THEM, BUT I'M NOT AS GOOD COMING UP WITH THEM AS MY COLLEAGUES. SO THE -- IT'S ONLY ONE OF MY DEFINITE INFORMATICIANS. IT'S A PLEASURE TO HEAR YOU TALK ABOUT PRECISION MEDICINE INITIATIVE AND ITS RELATIONS TO INFORMATICS, ERIC, IT'S A REAL PLEASURE, I INVITE YOU TO FEEL FREE TO INTERJECT IN THIS PROCESS, OF COURSE A DISCUSSION. THE REPORT WILL BE MADE PUBLIC ONE TO THREE TIME POINT UNFORTUNATELY I'M UNDER A BIT OF A GAG ORDER. I'LL RELATE PRINCIPLES WE DISCUSS IN OUR ON GOING DISCUSSIONS BOTH PUBLIC AND SOME OF THE EVALUATIONS WE MADE IN PRIVATE AND TALK TOWARDS THE IMPORTANT VIEWPOINTS WE DISCUSSED. SO THIS WAS ANNOUNCED FIRST STATE OF THE UNION I DON'T KNOW ON JANUARY 20th AND TEN DAYS LATER -- ANNOUNCE THE PRECISION MEDICINE INITIATIVE AND AS A GOAL WITH TWO PARTS, ONE TO TREAT CANCER AND ANOTHER TO FORM THE NATIONAL RESEARCH COHORT WITH A MILLION INDIVIDUALS TO REALLY ACCELERATE MEDICAL SCIENCE IN THIS COUNTRY. AND THE WORLD. AND HIS PROPOSED BUDGET FOR 2016 INCLUDED $215 MILLION AND THAT IS MOST OF WHICH IS GOING TO THE NIH BUT THEN ALSO COMPONENTS FOR THE FDA AND ONC REALLY THINKING OF THIS AS AN INTERAGENCY INITIATIVE AND THERE'S BEEN A LOT OF INTERDIGITATION THROUGHOUT THE ENTIRE PROCESS WITH PARTICIPATION IN THE WORKING GROUPS AS WELL AS JUST A LOT OF ONGOING DISCUSSIONS THROUGH THESE LAST FEW MONTHS. THE BREAK DOWN TO THE NIH INCLUDES 70 MILLION FOR PMI ONCOLOGY OR CANCER COMPONENT AND 130 MILLION TO START THIS NATIONAL RESEARCH COHORT. AND A LITTLE BIT MORE ABOUT THE VISION AND THE TWO INITIATIVES, THE VISION IS FOR A BROAD BASE RESEARCH PROGRAM THAT WILL GO ON IN TIME AND TO TEST THE VIGOROUSLY AND GUIDE EVIDENCE, THE EVIDENCE BASE NEEDED TO GUIDE CLINICAL PRACTICE. PRECISION MEDICINE INITIATIVE FOR ONCOLOGY WHICH I WILL NOT SPEAK MUCH ABOUT, REALLY ALMOST AT ALL TODAY. BECAUSE I KNOW YOU HEARD ABOUT THAT AT YOUR LAST MEETING. REALLY TO APPLY PRECISION MEDICINE TO CANCER, REALLY TO EXTENSION OF ONE OF THESE ISSUES EXTENSION OF EXISTING NCI MATCH PROTOCOL AND I KNOW YOU HEARD A LOT ABOUT THEM AT THE LAST BOARD OF REGENTS MEETING. THE LAST PMI COHORT PROGRAM WE CALL IT TO GENERATE A LOT OF INDIVIDUALS THAT WILL BROADLY SHARE HEALTH RELATED DATA AS A NATIONALLY ACCESSIBLE AND NATIONALLY REPRESENTATIVE COHORT WITHOUT FOCUSING ON WHAT THAT REPRESENTATION IN TERMS OF STATISTICAL TERM WOULD LOOK LIKE. THIS SHOWS MEMBERS OF THE WORKING GROUP NAMED IN MARCH. I DON'T REMEMBER, IT FEELS LIKE A LONG TIME AND IT'S SHARED BY CLIFTON, ALSO A MEMBER OF THE ADVISORY COMMITTEE TO THE DIRECTOR. AND PATRICK LAKE ADVOCATE ENGAGEMENT EXPERT AT DUKE, CATHY HUDSON I'M SURE MANY OF YOU KNOW IS A DEPUTY DIRECTOR OF NIH. AND THEN MEMBERS YOU CAN SEER RICK AND I ARE ON THIS COMMITTEE. WHAT YOU SEE HERE IS A DIVERSE REPRESENTATION FROM PRIVATE SECTOR, PATIENT ADVOCATES, AND ACADEMICS. THIS INCLUDES INDIVIDUALS WHO WORK WITH UNDER-REPRESENTED POPULATIONS, AND NATIVE AMERICAN POPULATIONS AS WELL AS JUST SORT OF HAVE A COMPLETE SPECTRUM WE THINK OF REPRESENTATION AND WITHIN INFORMATICS COMPONENTS OF THE THIS WE HAVE A DIVERSITY OF REPRESENTATION OF THE KINDS OF NETWORKS THAT WE HAVE BEEN EXPOSED TO. IN ADDITION, WE HAD WORKSHOPS, FOUR OPEN TO THE PUBLIC AND THEN HAD DISCUSSION GROUP AFTERWARDS CLOSED TO THE WORKING GROUP AND THEN A NUMBER OF TELECONFERENCES AND WEBINARS AND THINGS LIKE THAT. THESE WORKSHOPS INCLUDED ONE IN PARTICULAR IN MAY VANDERBILT, INDIVIDUAL DATA AND HOW WE WOULD DESIGN SUCH NETWORKS AND HOW WE SHARE DATA, WE TALKED ABOUT DIFFERENT MODELS THERE. AND LOOKED AT SOME OF THE SUCCESSES AND LESS SUCCESSFUL ENTERPRISES. AND THEN IN JULY TALK ABOUT INTEL AND MOBILE SENSOR TECHNOLOGIES, AND WHERE WE CAN INTEGRATE THOSE KINDS OF DATA AS WELL AND WHAT THOSE DATA SOURCES ARE ALL GUIDED BY THE FIST WORKSHOP IN APRIL WHICH ARE WHAT ARE SOME OF THE OPPORTUNITIES WE CAN USE A LARGE COHORT FOR. ALSO TWO NIH BASE REQUEST FOR INFORMATION LOOKING AT EXISTING BIOBANKS THAT ARE OUT THERE AND STRATEGIES WE CAN PURSUE TO ENGAGE PATIENTS AND WHO -- THERE'S A SURVEY DONE BY THE F NIH PUBLIC PERCEPTION AND DESIRE TO PARTICIPATE SUCH A COHORT WHAT THEY VALUE AND WHITE HOUSE PRIVACY AND TRUST PRINCIPLES THAT'S THE COORDINATED AMONG NUMBER OF AGENCIES AS WELL. THE WORKING GROUP HAS A CHARGE TO DEVELOP A VISION FOR THE PMICP, ANOTHER ACRONYM AND HOW WE GET A NATIONAL CORPS OF A MILLION PEOPLE THERE WOULD BE VOLUNTEERS TO SHARE DATA, WE LOOK AT DIFFERENT COHORT MODELS HOW TO START FROM SCRATCH, LEVERAGE EXISTING MODELS HOW TO CAPTURE THE DIVERSITY OF THE UNITED STATES. ACROSS NUMBER OF DIFFERENT WAYS YOU CAN DEFINE DIVERSITY, SOCIO ECONOMIC, GEOGRAPHIC, ETHNIC, WHAT DATA TYPES TO INCLUDE, ANOTHER IMPORTANT COMPONENT. AND THESE ARE SOME OF THE OPPORTUNITIES WE IDENTIFIED FOR THE COHORT AND MOST OFTEN -- ALL THIS INFORMATION IS EXPANDED UPON IN THE DOCUMENTS RELEASED TODAY. THINKING ABOUT QUANTITATIVE RISK ESTIMATES ENVIRONMENTAL EXPOSURES GENETIC FACTORS GENE ENVIRONMENT INTERACTIONS, SUCH LARGE COHORT ALLOW US TO DO SOME OF THOSE THINGS NOT PREVIOUSLY AVAILABLE. PHARMACOGENOMICS IS AN EARLY TARGET SO DRUG EXPOSURES AND RESPONSE. LOOKING AT BIOMARKER IDENTIFICATION. SO GENOMICS IS ONE TYPE OF BIOMARKER BUT THERE ARE MANY OTHER TYPES OF BIOMARKERS WE CAN LOOK AT. THE FACT THAT THIS IS COHORT IS PERSPECTIVELY IDENTIFIABLE ALLOWS US TO LOOK AT THE EFFECT OF EXPOSURES OVER TIME ON DISEASESNESS, DRUG RESPONSE, THINKING ABOUT DISEASE CLASSIFICATION, RECLASSIFICATION, INSTITUTE OF MEDICINE REPORT ON PRECISION MEDICINE COUPLE OF YEARS AGO THAT TALKED ABOUT TAX ON TAXONOMY CENTURIES OF E MEASUREMENT ANECDOTE AND HOW MORE DETAILED REFINED CLASSIFICATION MAYBE POSSIBLE WITH MOLECULAR DATA. MOBILE TECHNOLOGY AND OTHER SENSOR TECHNOLOGY IS SOMETHING YOU SEE A BIG POSSIBLE WIN HERE. AVAILABILITY TO INVESTIGATE LOSS OF FUNCTION MUTATIONS REQUIRED REALLY LARGE COHORTS MANY GENES HAVE THEM, SOME GENES WE DO NOT TOLERATE THEM BUT MANY WE MIGHT. MANY IN FACT WE DO. AND WE WANTED TO RETHINK THE MODEL ENGAGEMENT WITH PARTICIPANTS NOT JUST SUBJECTS, PARTNERS IN THE RESEARCH PROJECT AND PROCESS, THINKING WHAT THAT MEANS IN TERMS OF SHARING DATA, IN TERMS OF GIVING ACCESS TO DATA AND REALLY RELYING IN WAYS TO GET US DATA, THEY'RE GOING TO HAVE TO INTERACT IN AN ONGOING WAY TO MAKE THIS SUCCESSFUL. HAVING THESE DATA ALLOW TO THINK ABOUT MORE TARGETED CLINICAL TRIALS FOR THERAPEUTICS AND OREGON INSTANCES OVER TIME AND WE BUILD THAT CAPABILITY BY BEING ABLE TO RECONTACT PARTICIPANTS. WHY NOW, DATA YOU KNOW IT USED TO BE EXPENSIVE TO SEQUENCE A GENOME, NOTS SO MUCH ANY MORE, THAT WE CAN DO FASTER. THAT SMART PHONE USE IS ACCELERATED AND IT'S TRUE ACROSS SOCIO ECONOMIC BOUNDARIES. IT IS EHR ADOPTION ACCELERATE DRAMATICALLY AND AMOUNT OF DATA ACCESSIBLE IN THE EHR INCREASES WHICH DOESN'T MEAN IT'S WITHOUT CHALLENGES. THE COMPUTING POWER IS INCREASED DRAMATICALLY AND THE KINDS OF TECHNIQUES WE HAVE TO MINE DATA IS INCREASED DRAMATICALLY AS WELL. SO ASSEMBLING THE COHORT, THE OF A MILLION PEOPLE WE WOULD LIKE TO HAVE AS MANY PEOPLE AS POSSIBLE. WE SET A MILLION AS A GOAL THAT WE THINK IS OBTAINABLE. AND IN -- BUT IF WE COLLECT MORE THAT'S GREAT. AND KEY ATTRIBUTES WE DISCUSS AT THE MEETINGS WERE THAT WE NEEDED TO HAVE LONGITUDINAL ACCESS TO HEALTH DATA. WE WANT TO LEVERAGE DATA WE CAN GET FOR FREE SO THAT IMPLICATES ELECTRONIC HEALTH RECORD DATA FREE IN QUOTE OF COURSE, NOT HAVING TO PURPOSEFULLY COLLECT SPECIFIC DATA. OBVIOUSLY WE WANT BIAS SPECIMENS WE NEED THE ABILITY TO RECONTACT INDIVIDUALS. THOSE ARE KEY TENANTS THAT WE VALUED AS A WORKING GROUP. WE DISCOVERED A NUMBER OF MODELS TONE ABLE ENROLLMENT AND WE HAVE SPECIFIC RECOMMENDATIONS ON THESE, I WON'T DISCUSS, BUT THE PRINCIPLE IS WE WANT EVERYONE TO PARTICIPATE IN THE US. AND WE REALLY ALSO FELT IT WAS IMPORTANT TO LEVERAGE THE EXISTING RELATIONSHIPS WITH HEALTHCARE PROVIDER ORGANIZATIONS BECAUSE WE KNOW THESE HAVE BEEN SUCCESSFUL IN THE PAST. MANY FAMILIAR WITH, MANY LARGE EHR RELATED BIOBANKS AND THESE COLLECTED HUNDREDS OF THOUSANDS OF PEOPLE MAYBE MORE LIMITED ENGAGEMENT HOMOS THAN WHAT WE HOPE FOR PMI BUT SHOW THE POSSIBILITY TO THIS RESEARCH WITH A LOT OF SUCCESS. WE LOOKED AT THE POSSIBLE DISEASES WE COULD STUDY. IF WE HAD A MILLION PEOPLE TO MOTIVATE AND THINK ABOUT SOME OF THE DIFFERENT KINDS OF MODELS, WE CAN PURSUE IN THIS COHORT. SO THESE DATA COME FROM BASICALLY SIMULATIONS FROM VANDERBILT, 200,000 INDIVIDUALS WHO AGREED TO BE SHARED USE EHR DATA FOR BIOLOGIC RESEARCH, REALLY DNA, PROJECT OUT TO AND COMPARE ACROSS THE SITES EMERGE NETWORK TO SHOW THESE ESTIMATES ARE CONSERVATIVE. AND ALL BIOBANK PEDIATRIC AND ADULT. PREVALENCE THINGS YOU CAN DO GENETIC RESEARCH ON AND LOTS OF OTHER STUDIES AND SET UP YOUR PHARMACOGENETIC STUDIES BECAUSE A PERCENTAGE WILL BE ON MEDICATION OF INTEREST. INCIDENT CASES ARE WHERE YOU START TO SEE IMPACT LOOKING AT EXPOSURE AND BIOMARKERS DETECTION, WE ACCRUED A NUMBER OF INDIVIDUALS AS EXPECTED WITH A NUMBER OF INDIVIDUALS, STATISTICAL THRESHOLDS 5 TO 10,000 PEOPLE TO DO A DISEASE GENE STUDY, STATISTICAL SIGNIFICANCE, 20,000 INDIVIDUALS TO LOOK AT GENE ENVIRONMENT INTERACTION OR PHARMACOGENETIC STUDY SO YOU CAN SEE IN THERE WE HAVE A LOT OF INDIVIDUALS FOR A NUMBER OF DISEASES. WE TALK DATA SOURCES THE PRINCIPLE TO THINK DIVERSE DATA CLASSES FROM THE BEGINNING AND MAYBE START WITH A MORE LIMITED SET WHICH I -- THAT I -- WE CAN COLLECT AT THE BEGINNING BUT EXPECT THAT TO GROW AND PLAN FOR MORE, WE REALLY WANT THIS TO TAP INTO THE RICH DETAIL OF DATA WE CAN GET AND NOT BE LIMITED TO MORE WHAT CORE DATA SET THAT CAN BE STRUCTURED. THAT HAS IMPLICATIONS HOW WE DESIGN THE COHORT. HERE ARE THE CATEGORIES. SELF-REPORT MEASURES ARE IMPORTANT ESPECIALLY DIET, SUBSTANCE ABUSE, SUBSTANCE USE. MOVEED ASSESSMENT THINGS LIKE THAT. -- MOOD ASSESSMENT, FAMILY HISTORY AND STRUCTURED AND UNSTRUCTURED CLINICAL DATA. WE SEPARATE THE TWO CLASSES BECAUSE STRUCTURE DATA WE CAN MORE QUICKLY MAYBE MOVE TOWARDS RAPID QUERYING ON STRUCTURE DATA, NEED MORE METHODS TO THINK ABOUT ACTUAL PROCESSING, PROCESS ING AND MORE TECHNIQUES BUT WE THINK ABOUT IMAGE DATA, WE TALK A LOT ABOUT REPORT BIOSPECIMENS, GLOBAL AND SENSOR TECHNOLOGIES, COMMERCIAL AND RESEARCH GRADE TECHNOLOGIES AVAILABLE AND BRINGING THESE FROM THE BEGINNING AND LEVERAGING WHAT'S PREVALENT FROM DAY ONE SO THAT COULD IMAGINE COMMERCIALLY AVAILABLE ACTIVITY MONITORS PEOPLE USE SMART PHONE USE, THEY DON'T EXIST MAIN ROBUST STANDARDS FOR THESE NOW. THAT'S INFORMATICS AREA TO EXPLORE, DEVELOPMENT HOW TO USE THIS DATA. ONCE YOU HAVE IT YOU CAN USE THAT DATA WITH SMART PHONES VIABILITIAL GEOSPATIAL DATA AND THEN WE TALK OTHER DATA LIKE SOCIAL NETWORKING AND THINGS LIKE THAT. THE FNIH STUDY FOUND THIS WAS THE CLASS OF DATA THAT INDIVIDUALS WERE THE MOST RETICENT TO SHARE. SO THAT INVOLVED SPECIAL CONSIDERATION AS WE THINK ABOUT INDIVIDUALS SHARING THAT. SOCIAL DETERMINANTS OF HEALTH, HEALTHCARE DATA, MOBILE TECHNOLOGY, THESE ARE ENVISIONED AS THINGS WE WANTS AVAILABLE TO COHORT. AND STARTING WITH A MORE LIMITED SET OF THOSE SPECIFIED IN THE REPORT OF WHAT THAT CORE DATA WOULD BE. AND THEN JUST TO GIVE EXAMPLE OF DATA WE HAVE AT VANDERBILT OF 260,000 INDIVIDUALS, WE CAN DEFINE A GROUP THAT MAYBE ALL PATIENTS IS ABOUT 65% AT ABOUT FIVE YEARS. BUT WE CAN DEFINE THE INDIVIDUALS WE CAN FOLLOW UP FOR 80, 90% AND FIGURE OUT WHO THOSE INDIVIDUALS ARE BY EASY PROGNOSTIC FACTORS. WE HAVE 22 MILLION IC 99 CODES FROM INDIVIDUALS. WE HAVE 141 MILLION LABS 159 MILLION DRUGS ON 2,629,000 INDIVIDUALS, AVERAGING 155 NOTES PHYSICIAN STAFF, STAFF PATIENT ALL THESE INTERACTIONS ARE COLLECTED. AND THE THINGS AMENABLE TO RESEARCH AND 2.3 MILLION RADIOLOGY TESTS, 10 RADIOLOGY TESTS PER PERSON. GENERAL PRINCIPLES PROVIDE DATA FOR US. ORGANIZATIONS WILL TOO. ORGANIZATIONS LIKE HEALTHCARE -- MOST EFFICIENT MODEL O TO GET DATA INTO THE COHORT. HAVING PARTICIPANT DATA CAN ENABLE VOLUNTEERS. THAT'S PARTICIPANTS PROVIDING THEIR OWN HEALTHCARE DATA FROM INSURANCE COMPANIES OR HEALTHCARE SYSTEMS THEY HAVE ENTERACTED WITH HOSPITALS, LOCAL PROVIDERS. AND MEANINGFUL USE STANDARDS. WE WANT TO -- WE TALKED ABOUT HOW TO ORGANIZE DATA. AND THE MODEL HAVING CORE DATA THAT COULD BE CENTRALLY AVAILABLE ROSE AS A TENANT. THAT WILL ENABLE FASTER QUERIES AND SCALE TO COHORT. BUT IT'S NOT SUFFICIENT BECAUSE WE MAY WANT TO KNOW ABOUT DETAILED AROUND DRUG REACTION, WE HAVE TO GO BACK TO THE ACTUAL NOTES IN THE PATIENT CHART. SO WE WANT TO PRESERVE THAT ABILITY TOO. WHICH IMPACTS OUR DATA MODEL. WE RECOGNIZE A LOT ABOUT HETEROGENEITY OF EHR DATA AND GETTING THE DATA A SINGLE PLACE AND THE ACD WORKING GROUP HAD A VARIETY OF INTIMATE EXPOSURES AND WORKING KNOWLEDGE OF DIFFERENT NETWORKS DIFFERENT TECHNOLOGIES, TO DO THESE KINDS OF DATA SHARING. AND WE CONSIDER THREE MAJOR MODELS, CENTRALIZED MODEL WHERE ALL DATA IS ONLY DATA YOU LOOK AT IS CENTRAL DATA STORE. PURELY FEDERATED APPROACH THAT HAS HUMAN FEDERATION HUMANS EXECUTE THE LOGIC OR A COMPUTERIZED FEDERATION THAT'S LIMITED TO SORT OF THINGS THAT CAN BE EXPRESSED IN COMMON STRUCTURED DATA. OR A BIT OF HYBRID. AS YOU MIGHT GUESS WE HIGHLY VALUE A HYBRID MODEL, THAT -- SORRY THE MIDDLE WITH EXISTING NUMBER OF EXISTING RESEARCH DATA CENTERS OR HEALTHCARE PROVIDER ORGANIZATIONS THAT CAN PROVIDE INDIVIDUALS TOWARD IN WITH COMMON LOGIC TO SHARE DATA, AT LEAST IN A CORE DATA MODEL OR WHAT CAN BE STRUCTURED, TOWARD CENTRALIZED DATA CENTER WHICH STORE SOME AMOUNT OF DATA CENTRALLY. THAT AMOUNT AMOUNT OF DATA CAN GROW OVER TIME. IF YOU LOOK AT THE FIGURE HERE, YOU IMAGINE YOU HAVE A NUMBER OF VOLUNTEERS, AT THE BOTTOM THAT WANT THE TO SHARE DATA YOU CAN HAVE IT BEING SHARED SOME COMPONENT OF IT AS CORE DATA, COMMON DATA MODEL, UP TO A COORDINATING CENTER. THAT WOULD ENABLE REACH QUERIES THAT CAN BE EXECUTED ENTIRELY WITHIN SOME DEGREE OF CORE DATA, TO BE DONE JUST AT COORDINATING CENTER GET SCALE HUNDREDS OF THOUSANDS OF THE QUERIES. WE NEED STANDARDS WHAT HA LOOKS LIKE AND LEVERAGE EXISTING STANDARDS WHERE POSSIBLE, MANY REPRESENTATIONS LIKE SNOW MED THAT CAN BE USED AND THEN THERE ARE EXISTING DATA MODEL STANDARDS OUT THERE TOO THAT GIVE US GOOD STARTS, IF NOT FULLY IMPLEMENTABLE TECHNOLOGY THAT WE WANT TO PURSUE. OVER TIME THAT EXPANDS, SOME BY COORDINATING CENTER AND SOME BY LOCAL SITES, GOING BACK TO PARTICIPANTS IN PMI OR GOING BACK TO HEALTHCARE CENTERS THAT PERSIST PMI. AND THEY CAN AUGMENT WHAT YOU CAN DO CENTRALLY. WE LOOK AT WHAT TYPES OF DATA WE NEED, BY DOING SO WE LOOKED AT 92 PHENOTYPE ALGORITHMS USED FOR GENOMIC DATA PUBLISHED IN PKB WHICH IS THE SITED THAT AGGREGATES THESE SORTS OF ALGORITHMS, 96% YOU CAN SEE THAT ICD CODES CODES FROM THE MOST COMMONLY USED ATTRIBUTE, AT 70% ALGORITHMS, MEDICATIONS WERE NUMBER TWO, AT 62% SOME DEGREE OF TEXT PROCESSING WAS USED IN 46% CPT CODES LABORATORY FILL OUT THE REST OF THE DATA TYPES THAT WE LOOKED AT 92 ALGORITHMS. THIS SAYS WE NEED MULTIPLE CLASSES OF DATA AND ALGORITHMS INCORPORATE DIFFERENT CLASSES AND EITHER LOGIC OR MACHINE LEARNING OR DECISION TREE TYPE APPROACHES TO GET ACCURATE ALGORITHMS. THIS GRAPH SHOWING LIEU AT ANY ONE CLASS OF DATA YOU CAN GET MAYBE REASONABLE PERFORMANCE AREA UNDER THE CURVE AND OPERATOR CURVE ANALYSIS ONE IS PERFECT. .5 IS RANDOM CHANCE. BUT WHEN YOU COMBINE TWO OR MORE COMPONENTS YOU CAN SEE A DOT, YOU GET BETTER PERFORMANCE, A SIMPLE COMBINATION AND MUCH MORE SOPHISTICATESSED COMBINATIONS AND IMPROVE PERFORMANCE FROM THERE. SO HOW DO YOU COLLECT EHR INFORMATION? IF YOU ARE A LARGE HEALTHCARE SYSTEM, YOU HAVE A LOT OF THESE DATA YOU PROBABLY HAVE A RESEARCH DATA WAREHOUSE AND YOU CAN RESTRUCTURE THESE DATA ESPECIALLY THE STRUCTURED DATA AND SEND IT FORWARD. WE CAN PUT CERTAIN STANDARDS AND TECHNOLOGY AND THAT WOULD PROBABLY BE HELPFUL TO STANDARDIZE THAT'S SPECIALLY WHEN YOU TAKE A LOT OF BROAD FOLKS DOING THIS. YOU STILL WANT TO PRESERVE THE ABILITY TO -- FOR LOCAL ACCESS TO THIS. FOR INDIVIDUAL PARTICIPANTS THOUGH WE NEED TECHNOLOGIES TO SHARE THESE DATA. THERE ARE STANDARDS THAT HAVE BEEN PUT FORTH WE BELIEVE AND WE TALKED A LOT ABOUT BLUE BUTTON BUT IT'S NOT REALLY ALL THE STANDARDS THAT WE NEED AROUND THEM DON'T EXIST YET. SO WE CAME UP WITH THIS IDEAS OF SING FOR SCIENCE. AND REALLY EVOLUTION THE VIEW DOWNLOAD AND TRANSMIT PROTOCOL DISCUSS MEANINGFUL USE TO OR THE BLUE BUTTON WHICH IS IMPLEMENTED VARIETY OF WAYS. BASED ON DIFFERENT PROVIDERS WHERE IT'S MORE LIMITED OR BROAD-BASED. WOULD BE A BROADER TOOL THAT LINK OR SYNC SOMEONE'S HEALTHCARE SYSTEM DATA TO POTENTIALLY RESEARCH ENTERPRISE BUT ALSO EXTEND TO ANY SORT OF CLINICAL ENTERPRISE. SO THAT COULD BE A PROTOCOL BY WHICH ONE INDIVIDUAL COULD BE HEALTH INFORMATION EXCHANGE AND THEY COULD BE HEALTH INFORMATION EXCHANGE AND SAY CONNECT MY DATA FROM HOSPITAL A OR HOSPITAL B FOR INSTANCE OR PROVIDER X. THEN THEY CAN BE THE ARBITER OF SHARING THAT DATA AMONG A NUMBER OF INDIVIDUALS AS WELL AS ENABLING NEW TYPES OF RESEARCH, THE IDEA WOULD BE THAT IT WOULD BE WHOLLY EHR DATA PER CIVIC HOPEFULLY MOVING THE STRUCTURED AND UNSTRUCTURED DATA AND WE WOULD NEED TO THINK ABOUT ELEMENTS METADATA TO SUPPORT THIS. THERE ARE PROJECTS ON GOING TO INFORM THIS, THE INTEROPERABILITY, I CAN'T REMEMBER WHAT R STANDS FOR RESOURCE, FIRE. AND ARGUE NAUGHTS PROJECT USING THAT FOR INSTANCE. WE VALUE HAVING SECURE COMPUTING ENVIRONMENT SO ENCLAVE BEING ONE EXAMPLE. DATA SECURITY AND PRIVACY, VERY IMPORTANT. WE EXPECT IT TO BE A TARGET AND APPROACHES TO PROTECT THAT AND DEIDENTIFICATION. WE FELT IN GENERAL WORK SHOULD BE DONE ON DEIDENTIFIED DATA EXCEPT WHEN THERE'S A USE CASE THAT NEEDS IDENTIFICATION. THAT INCLUDES WHOLE LANGUAGE NATURAL LANGUAGE NARRATIVE TEXT THOUGH REDUCING RISK OF IDENTIFICATION WHEN IT'S NOT NECESSARY IMPORTANT. ROLE BASED ACCESS OF COURSE IS IMPORTANT. AND ESPECIALLY RESTRICTING AND IMPROVING INDIVIDUAL LEVEL ACCESS TO DATA, PARTICIPANTS SHOULD HAVE ACCESS TO DATA, THAT'S AN IMPORTANT PARADIGM HERE. THINKING ABOUT THE IMPLICATIONS OF THAT IS IMPORTANT TOO, AN AREA THAT WILL NEED TO BE THOUGHT ABOUT, ACTIVE AREA OF RESEARCH. STARTING WITH ACCESS TO CLINICAL DATA, EVERYONE SHOULD HAVE ACCESS TO CLINICAL DATA. THINK GENOMIC DATA. SOME OF THAT RETURN OF ACCESS BECOMES MORE NUANCED. WE SPEND TIME TALKING ABOUT THAT AND WE HAVE A LOT OF LANGUAGE IN THE REPORT ABOUT THAT. BUT TALK A NEW CLASS OF DATA TO BRING IN, MOBILE TECHNOLOGIES AND HOW WE CAN PROMOTE STANDARDS IN THIS AREA, MUCH DATA STRUCTURED BUT IN ARBITRARY WAYS AND NOT X COMPATIBLE WAYS CURRENTLY. IMPLEMENT DATA EARLY, PLAN NOR DATA AND MORE DATA TYPES AS WE MOVE ALONG. THANK YOU OBVIOUSLY O WORKING GROUP MEMBERS ERIC AND OTHERS WE HAVE LOTS OF IN THE WORKSHOPS AND THE PLANNING TEAM AN NIH STAFF WHICH THE SECRETARY COORDINATOR FOR ALL THESE EFFORTS AND INTEL FOR HOSTING ONE OF THESE. I WON'T THANK OURSELVES. THANK YOU VERY MUCH FOR YOUR TIME. [APPLAUSE] >> I WILL LET YOU, YOU GUYS HAVE (INAUDIBLE) >> I THINK THIS IS INCREDIBLY EXCITING AND I THANK YOU BOTH FOR ALL YOU'RE DOING HERE. SOMEBODY ASKED ME BECAUSE I WAS MENTIONING THE GSA SO MUCH AS A GSA AND I AM NOT. BUT MAYBE SHE WAS GIRL SCOUT. SO I'M AT UNIVERSITY OF ARIZONA, USED TO BE AT NIH. WE HAVE BEEN DOING A PIECE OF THIS, AS THE GSA LINKING 12 DIFFERENT ENVIRONMENTAL ATTRIBUTES IN REAL TIME WITH HUMAN HEALTH AND WELL BEING OUTCOMES PHYSIOLOGICAL BEHAVIORAL, SMART PHONES. AND MY QUESTION IS, HOW ARE YOU GOING TO MEASURE A MILLION PEOPLE WITH ONLY $200 MILLION? >> FISCAL YEAR 2016 BUDGET. >> IT CAUSED A LOT OF MONEY TO DEVELOP THE ALGORITHMS. WE HAVE LIKE 44 ON OUR TEAM, WE HAVE ANALYTICS WE HAVE BIOENGINEER PHYSICAL PROCESSING PEOPLE, WE HAVE PSYCHOLOGY, SMART PHONE -- A HUGE ENDEAVOR. >> IT IS HUGE. WE WANT -- I THINK A LOT OF PARTICULAR PROJECTS MAY ACTUALLY INVOLVE PARTICULAR -- PARTICULAR GRANTS THAT AUGMENT DATA SOURCES INTO THE COHORT. IF YOU WANT TO BUY A SMART PHONE YOU NEED FIGURE OUT A GREAT RELATIONSHIP WITH A COMPANY THAT WANTS TO PROVIDE SMART PHONES OR YOU HAVE TO FIGURE A WAY TO PAY FOR IT. BUT AGAIN, THESE ARE THE INITIAL FUNDS INTO THE COHORT AND AS SCIENTIFIC OPPORTUNITIES ARISE YOU CAN IMAGINE OTHER WAYS OF FUNDING DIFFERENT THINGS. >> THAT'S WRIT'S GOING, IT HAS TO BE PUBLIC PARTNERSHIP. I DON'T THINK IT'S POSSIBLE -- I MEAN, WE ACTUALLY HAVE A COMPANY THAT WE'RE WORKING WITH IN THE BAY AREA THAT HAS DEVELOP METHODS TO MEASURE THESE REAL TIME ENVIRONMENTAL ATTRIBUTES. LINKING THAT PLATFORM TO THE MOST UP TO DATE VARIABILITY BEHAVIOR SLEEP CENTRAL MONITOR DEVELOPED IN GERMANY WHICH DOESN'T HAVE FDA APPROVAL YET, IT IS COMPLICATED. >> WE DID BENCH MASHING ABOUT WHAT OTHER NATIONS ARE INVESTING, MORE EQUIVALENT (INAUDIBLE) INITIATIVE AND WHY LARGE BUDGET ITEM WE'RE REQUESTING AT THIS POINT WE'RE TRYING TO BUILD THAT PATIENTS AMONG POLICY MEMBERS AMONG BOTH SIDES OF THE AISLE, THIS IS GOING TO BE A LONGER TERM NATIONAL INVESTMENT AND THIS IS THE DOWN PAYMENT IF YOU DON'T WANT TO DO IT DON'T DO THE DOWN PAYMENT. IT WILL BE A JOURNEY. >> WHEN WE TAKE IT TOGETHER WITH THE REAL TIME MOLECULAR BIOMARKERS WHICH IS ANOTHER SHOT, IT'S GREAT. IT'S GREAT BECAUSE (INAUDIBLE) DEVELOPING. >> >> VERY ROUGH IDEA OF TIME LINE AND MILESTONES AS YOU SEE THEM THERE. >> ALSO DON'T KNOW HOW MUCH I CAN REALLY SAY. I DON'T KNOW HOW MUCH -- SO THERE ARE SOME TIME LINES THAT ARE BEING DISCUSSED ON HOW THIS HAPPENS. I CAN TELL YOU REPORTERS FOR AT LEAST TODAY I AS A MEMBER OF VANDERBILT UNIVERSITY HAVE NOT BEEN PART OF RFA DISCUSSION OR THINGS LIKE THAT SO I DON'T KNOW THE SPECIFIC TIME LINE THERE. I KNOW PRIORITY SAYINGS FOR COMPONENTS AND STRUCTURES IN THE REPORT TOWARDS BUILDING AND STARTING RECRUITMENT AND ENROLLMENT PROCESS. SO THE INDIVIDUAL -- I CAN ALSO TELL YOU IF YOU LOOK AT ALL THE OTHER EXISTING COHORTS AND HOW LONG IT TOOK TO ENROLL THEM, YOU'RE PROBABLY LOOKING AT 4 TO 5 YEARS TO GET TO A MILLION. O IF YOU'RE CONSENTING INDIVIDUALS INTO A NEW COHORT TO DO NEW KINDS OF RESEARCH. ASSUMING YOU GET LOTS OF PEOPLE ACROSS A NUMBER OF DIFFERENT SITES. NOW YOU DON'T HAVE TO WAIT EVERYONE IS ENROLLED TO START DOING RESEARCH STUDIES BUT YOU DO HAVE TO GET A CERTAIN AMOUNT OF INFRASTRUCTURE IN PLACE BEFORE YOU START ENROLLING LOTS OF PEOPLE. ESPECIALLY IF YOU'RE PARTNERING WITH HEALTHCARE PROVIDER ORGANIZATIONS. ERIC, WHAT -- Q. THAT'S AT CASE -- I WON'T GET SPECIFIC -- ANOTHER CASE WHERE WE LEARN FROM MISSTACKS OF OTHERS INCLUDING OTHERS TO SORT OF UNDERSTAND, PARTICULARLY WHEN YOU HAVE SOMETHING SORT OF SO -- THAT'S BECOME SO PUBLIC AND YOU WANT THE PUBLIC TO BE EXCITED AND YOU WANT TO JOIN IN, SO THERE'S THIS PRESSURE TO GET STARTED AT THE SAME TIME LIKE BUILDING BOTH THE COHORT RECRUITMENT INFRASTRUCTURE AND SCIENTIFIC INFRASTRUCTURE TO DO IT. AND WOULD SERVE OURSELVES WELL. >> WE INVITE EVERYBODY TO PLAY AND WE'RE DEALING HERE WITH EXPECTATION OF WE TAKE BIOSPECIMENS, CONTRIBUTED TO THIS. I MEAN, YOU DO NEED TO HAVE A PLACE READY TAKE THOSE AND PROCESS THOSE BEFORE YOU SAY OKAY, COME. ONE OF THE THINGS THAT -- YOU THINK THAT THE EXPECTATION IS THAT IN ORDER TO PLAY YOU WILL HAVE TO NOT ONLY AGREE TO THE ORIGINAL THING, NOT ONLY AGREE TO WIRE ELECTRONIC HEALTH DATA TO BE ENROLLED AND RECONTACTED BUT YOU WILL HAVE TO AGREE TO HAVE MONITORING DEVICES. >> THOSE PROPERTIES ARE VERY SPECIFIC. WE ADDRESSED THE ISSUES >> ONE OF THE PRINCIPLES, THE PRINCIPLE AROUND DIVERSITY OF DATA TYPES SO THERE'S A PRINCIPLE AROUND DIVERSITY OF PEOPLE, IN ALL SENSE OF THE WORD. WE'RE WELL WARE HAVE ACCESS TO -- WARE OF CERTAIN ACCESS TO TECHNOLOGIES AND TRAINING AND SOME USE IT THAT WAY SO I MEAN ALL THOSE SHIRES HAVE SPECIFIC RECOMMENDATIONS. >> OBVIOUSLY (INAUDIBLE) JUST GENERALLY THINGS ABOUT IT I CAN IMAGINE THAT THERE WOULD BE PEOPLE SAY SURE THIS IS A GREAT IDEA, OR I WILL DO THIS, I DON'T MIND GOING IN, I DON'T MIND TRANSFERRING, I DON'T MIND EXTRA SAMPLE, I DON'T MIND THE FACT YOU CAN GET BACK IN TOUCH WITH ME FROM TIME TO TIME. I'M NOT GOING TO CARRY ANYTHING. I TURN THAT UP ON MY PHONE ALREADY. AND I DON'T REALLY WANT PEOPLE TO KNOW WHEN I'M UPSTAIRS OR DOWN STAIR OR WHEREVER I AM. I WONDER WHETHER THERE WAS A DISCUSSION AROUND THAT INFORMATION AND I CAN UNDERSTAND IT FOR A PROJECT LIKE THIS SO IMPORTANT. THAT THAT MIGHT BE A CUT FACTOR AS TO WHETHER -- >> >> WE HAVE FOLKS TO TAKE CERTAIN COMMUNITIES AND VERY DIFFERENT RECRUITMENT STRATEGIES, EDUCATIONAL STRATEGIES, MAINTENANCE STRATEGIES KEEPING THE COHORT. IT IS NOT ONE SIZE FITS ALL, THIS IS PARTLY WHY (INAUDIBLE) >> WE GRAPPLE WITH THIS WITH THE STUDY (INAUDIBLE) GS ACTION STUDIES BUT THIS IS SOMETHING THAT TOOK US A LONG TIME TO SET UP AND WE REALIZED WE HAD TO DO THAT BEFORE WE RECRUITED THE FIRST SUBJECT PERSON. PARTNER. AND SO IT DOES REQUIRE EDUCATION FROM THE BEGINNING, IT DOES REQUIRE -- ONE OF THE CRITERIA BECAUSE FEDERAL WORK FORCE, WE COULD NOT TRACK -- WE COULD NOT USE LOCATION TRACK. SO WE HAD TO DEVELOP ALGORITHMS TO FIGURE OUT A WAY TO LINK THOSE DIFFERENT DATA SETS WITHOUT THE -- IT'S EASIER TO DO IF YOU HAVE LOCATION TRACKS BUT -- WE FIGURED THAT OUT, THAT WAS PART OF IT. >> FEDERAL (INAUDIBLE) >> ACTUAL PROHIBITIONS BUT THE JUST LIKE YOU SAID -- >> SHE'S DEALING WITH GSA BUILDINGS BECAUSE THEY'RE TRYING TO HAVE HEALTHY BUILDINGS AND EVERYONE IN THE GSA BUILDING IS FEDERAL EMPLOYEE. MOST OF THEM. THAT PUTS ANOTHER CONSTRUCT AROUND IT BECAUSE SIMPLY THE FEDERAL GOVERNMENT ORDERED TO DO SOMETHING. (OVERLAPPING SPEAKERS) >> THE POINT IS THERE ARE WAYS AROUND ALL THESE ALGORITHMS SO YOU CAN FIGURE WHICH IS MORE COMPLICATED, COSTS MONEY WHICH IS WHAT I WAS SAYING. AND IT MAKES THE LEAD TIMES WAY, WAY LONGER. BECAUSE YOU HAVE TO GET ALL YOUR DUCKS IN A ROW BEFORE YOU START. >> IS THERE A COMMENSURATE LEVEL OF EFFORT TO SPELL OUT THE RIGHT DATA MODEL LOOKING TOWARDS THE FUTURE? TENS OF MILLIONS OF VETERANS SEEN IN DHA AND QUICKLY BROUGHT TO ITS KNEES WITH THE KIND OF QUESTIONS THAT PEOPLE ARE INTERROGATING IT WITH THAT WE HAVE (INAUDIBLE) 1970s AND 80s, DATA SET UP FOR THIS STUFF. THAT TO ME SEEMS REALLY CRITICAL AND POSSIBLY SOMETHING UNIQUELY SUITED FOR GOVERNMENT TO HELP FACILITATE. >> THERE WAS DISCUSSION OF THAT AND THINKING WE NEED TO SCALE THIS BEYOND WHAT WE ENVISION TODAY. I HAVE GONE OUR SYSTEM (INAUDIBLE) WE HAVE A MILLION INDIVIDUALS, BUT IT IS -- AND WHEN YOU THINK ABOUT IMAGING DATA, SEQUENCE DATA, THINK ABOUT ALL THE SENSOR DATA THAT COULD GET VERY TIME -- TIGHT TIME FRAME. THEY'RE PROBABLY ARE NOT NECESSARILY ANY ACADEMIC FACILITIES OUT THERE THAT CAN CURRENTLY HANDLE THESE CLASSES OF DATA AND OF THE SCALE FOR INSTANCE THAT WE WOULD NEED TO. SO YEAH, WE TALKED ABOUT THAT KIND OF THING. IT IS OF COURSE MORE GENERAL TERM THAN THE REPORT BUT WE'RE NOVEL WAYS, THE COMMITTEE VALUED MORE NOVEL WAYS OF LOOKING AT SUCH QUESTIONS. >> DISCUSSIONS GOING ON WITH DOE AND OTHERS USING THEIR KNOWLEDGE AND OTHERS TO HELP FIGURE THESE THINGS OUT. JUST TO GIVE THEM SENSE OF SCALE, WE DID SOME CALCULATIONS FOR SAKE OF -- IF WE SEQUENCED ALL THE 1.65 MILLION NEW CANCER PATIENTS ONES THAT CAME IN, AND ADDED THE OTHER CLINICAL DATA TYPES THAT WERE ACTUALLY TALKING ABOUT YOU GENERATE MORE DATA WHICH IS 400,000 TIME DIGITIZATION OF LIBRARY OF CONGRESS, THE DIGITIZATION OF ALL HUMAN SPEECH EVER DONE. SO IT WOULD TAKE $58 MILLION OF ELECTRICITY A YEAR, 15 MILLION SERVER (INAUDIBLE) SO AND YOU KNOW YOU START ADDING THINGS LIKE MOLECULAR DIAGNOSTIC MODELS. >> THERE'S ANOTHER ISSUE, REMINDED ABOUT COMPUTING (INAUDIBLE) OBITUARY TODAY YOU CAN DO THE HEAVY LIFTING COMPUTATIONALLY OR MATHEMATICALLY BUT AT THE END TIMES IT'S THE SIMPLE STUFF, LONGITUDINAL STUDY, TIME FRAME EPIDEMIOLOGISTS, JUST ENSURING FOLLOW-UP BECAUSE DIFFERENTIAL LOSS TO FOLLOW-UP BIAS RESULTS HUGELY JUST MAKING SURE THAT PATIENT IDENTIFIERS ARE KEPT CONSISTENTLY OVER TIME. THOSE ARE THINGS THAT ARE OUR BIGGEST CHALLENGES DATA MODELS AND IT'S REALLY (INAUDIBLE). >> VERY IMPORTANT. DONE THOSE KINDS OF THINGS, RECOGNIZE THE CHALLENGES BUT THIS IDEA OF A BALANCE OF WANTING TO REPRESENT, THIS IS NOW BEYOND THAT BECAUSE WE'RE HANDING OFF A REPORT. BUT WANTING TO BALANCE THE CHALLENGES GETTING DIVERSITY IN ALL FORMS. AND SORT OF WHERE YOU HAVE THE MOST FOLLOW-UP AND WHERE YOU SORT OF HAVE THE MOST CONTAIN MENT MAYBE IN TERMS OF WITHIN A HEALTHCARE SYSTEM. >> TIME FOR ONE MORE QUESTION. >> I AM FROM THE FDA (INAUDIBLE) YESTERDAY UNIQUE IDENTIFICATION AND ONE LAST THING I SAID WAS THAT I WAS (INAUDIBLE) ASK FOR THEIR UDIs, FOR THE IMPLANTABLE (INAUDIBLE) TO THE POINT THAT REGULATION OR IMPLANTABLE DEVICE AGENT IS THIS MONTH, I DON'T SEE ANYTHING PMI TALKED ABOUT DEVICES OR CAPTURING THAT DATA, IT'S A GREAT OPPORTUNITY AT THIS POINT IS GOING TO ROLL OUT TO TO INCLUDE THAT (INAUDIBLE) IN THIS. >> WE'RE AT BREAK AND ANYONE WHO HAS QUESTIONS PLEASE FEEL FREE, WE HAVE OUR TWO EXPERTS HERE. >> IF YOU READ THE REPORT, THIS IS UNBELIEVABLE REPORT. HE REALLY DID. HE TAUGHT MOST OF IT. I FELT LIKE HE GAVE A LOT OF SPEECH (INAUDIBLE). [APPLAUSE] >> SO BREAK FOR 15 MINUTES AND WE'RE BACK AT 10:45. >> SO OUR NEXT PRESENTATION IS BY DR. MICHAEL FELDGARDEN FROM THE NCBI. HE WILL BE TALKING ABOUT THE PATHOGENIC SEQUENCES UPDATE. >> THANK YOU FOR INVITING ME AND GIVING ME THIS OPPORTUNITY TO PRESENT THE WORK WE'RE DOING ON ANTI-MICROBIAL RESISTANCE INFORMATICS. ONE THING THAT WAS ALLUDED TO THE LAST DISCUSSION BY MICROBIAL GENOMICS IS COMMONLY USE AS AN EPIDEEM ROUGE LOGICAL TOOL -- EPIDEMIOLOGICAL TOOL. DO TWO ISOLATE BELONG O THE SAME OUTBREAK. HAVE WE SEEN THE ISOLATE BEFORE IN ANOTHER PART OF THE COUNTRY SO IF I FIND SOMETHING IN PEANUT BUTTER IN CALIFORNIA SALMONELLA IN CALIFORNIA AND ONE IN NEW YORK IS IT THE SAME PROCESSING PLANT OR JUST BAD LUCK FOR TWO SEPARATE PLACES. ALSO ASK IF WE SEE SOMETHING IN A CONTAMINATED PLANT, IS IT SHOWING UP IN THE LOCAL HOSPITAL. THOSE KINDS OF QUESTIONS. ONE ADVANTAGE OF THIS IS IT IS FASTER AND PROVIDES MORE RESOLUTION. BUT THE OTHER THINGS THAT PROVIDES MUCH MORE INFORMATION. MICROBIAL GENOMICS DOES ABOUT THE ORGANISM ITSELF AND ALLOWS US TO ANSWER A VERY IMPORTANT QUESTION ANTI-MICROBIAL RESISTANCE. I'M GLAD YOU'RE LAUGHING. I HAVE BEEN USING THIS SLIDE FOR A LONG TIME AND WHEN I FIRST STARTED USING IT I WOULD SAY WE'RE ONE DRUG CLASS AWAY FROM UNTREATABLE ORGANISMS. I CANNOT SAY THAT ANY MORE. WE NOW HAVE UNTREATABLE ORGANISMS. THIS IS A SERIOUS PROBLEM. AND ONE THING SO WE ASK OURSELVES HOW CAN NCBI COMBAT AMR. THE BROAD OVERVIEW SENSE THERE'S SEVERAL THINGS WE CAN DO, WE CAN PROVIDE DATABASES RESOURCES WE DO A LOT OF BAD, WE CAN IMPROVE IDENTIFICATION AND ANNOTATION OF RESISTANCE GENES, THIS IS BEING USED BOTH IN OUR PATHOGENERAL DETECTION PIPELINE, -- PATHOGEN DETECTION PIPELINE,LY TALK MORE ABOUT THAT LATER AND ALSO USENING OUR OTHER ANNOTATION PRODUCTS SO PEOPLE SEND GENOMES AND SAY CAN YOU TELL US WHAT GENES ARE THERE. THIS WILL -- WE'RE GOING TO BE USING THIS FOR THAT. FINALLY SOMEBODY MENTIONED EARLIER A BIT ABOUT STANDARDIZATION, WE CAN HELP STANDARDIZE AND IMPROVE RESISTANCE GENE NOMENCLATURE SO BIOLOGISTS ACROSS DISCIPLINES HAVE A COMMON LANGUAGE THEY CAN USE. EARLIER THIS YEAR THE WHITE HOUSE RELEASED THE NATIONAL ACTION PLAN FOR COMBATING ANTIBIOTIC RESISTANCE BACTERIA AND ONE OF THE CHARGES WAS TO DEVELOP MAINTAIN NATIONAL SEQUENCE DATABASE OF PATHOGENS. IF THIS SOUNDS SOMETHING LIKE GEN BANK YOU'RE RIGHT. THIS IS AN OBVIOUS FIT FOR KINDS OF WORK WE DO. THE CARD REPORT IT WAS UNDERSTAND THE RELATIONSHIP BETWEEN ANTIBIOTIC RESISTANCE PHENOTYPES, OTHER WHAT DRUGS YOU'RE RESISTANT OR SENSITIVE TO AND UNDERLYING GENOME. THE GENOTYPES. WE ALSO SAW THIS AS AN OPPORTUNITY TO BUILD AN INFRASTRUCTURE BASS FOR COLLECTING AND PRESENTING ANTIOBIOGRAMS THAT PATTERN OF RESISTANCE AND SENSITIVE THETY FOR A GIVEN BACTERIUM ASO I SHALLIATED WITH SUBMITTED BACTERIAL GENOME. THAT'S CLINICAL DATA OUT THERE, NOT PART OF GOVERNMENT SURVEILLANCE PROGRAMS OR NIH GRANTS, WE WANT TO BE ABLE TO GET THAT TOO. AND BEFORE I WAS NCBI I WAS AT THAT TIME BROAD INSTITUTE WHERE I SEQUENCED THOUSANDS OF GENOMESES AND I CAN SAY THIS IS SOMEWHAT AUTHORITATIVELY, WE DON'T NEED MORE BACTERIAL GENOMES. I DID ENOUGH OF THAT. WHAT WE NEED IS GENOMES THAT HAVE CRITICAL METADATA. THINGS LIKE RESISTANCE, THINGS LIKE WHAT IS THE HOST. THAT IS WHAT WE NEED TO GET A HANDLE ON THE ANTI-MICROBIAL RESISTANCE PROJECT. TO START DESCRIBING SOME OF THE THINGS THAT WE HAVE DONE, WE HAVE AN ANTIOBIOGRAM -- WE HAVE -- PEOPLE ARE ABLE TO SUBMIT ANTIOBIOGRAMS THIS IS JUST A SCREEN SHOT TO SHOW YOU WE ARE OPEN FOR BUSINESS. IN THREE MONTHS WE RECEIVED 700 ANTI-BUY GRAMS WITH GENOMIC DATA FROM A VARIETY OF SOURCES, SOME ARE THE GENOME SEQUENCING CENTERS FUNDED BY IONIA IDEA. SOME ARE PRIVATE INSTITUTIONS ALONG WITH GOVERNMENT AGENCIES THAT ARE PARTNERS AND WE WORK WITH DIFFERENT GOVERNMENT AGENCIES. PEOPLE HAVE THE ABILITY TO FIND GENOMES THAT HAVE THIS ANTIOBIOGRAM DATA THEY WANT ANALYSIS. JUST VERY BRIEFLY TOY EXAMPLE TALK ABOUT ANTIOBIOGRAMS FIELD, IN SHORT WHAT WE TRY TO DO IS GET THE INFORMATION WE NEED WITHOUT MAKING TOO OBNOXIOUS FOR PEOPLE TO SUBMIT BECAUSE WE'RE ASKING PEOPLE TO GIVE SOMETHING AND THEY MAY NOT GET SOMETHING DIRECTLY BACK IN RETURN RIGHT AWAY. ONE THING WE HAVE DONE TO HELP MAKE THIS SIMPLER IS UPDATED DICTIONARIES THAT WE HAVE DEVELOPED WITH OUTEXPERTS FROM THE AMERICAN SOCIETY OF MICROBIOLOGY AS WELL AS CLSI LABORATORY STANDARDS GROUP AND THIS ALLOWS US TO STANDARDIZE INPUTS AS WELL AS MINIMIZE DATA DRIFT SO OVER THE YEARS PEOPLE AREN'T CALLING THINGS DIFFERENCE THINGS, IT'S HARD TO PULL DATA OUT. THE KINDS OF THINGS WE COLLECT ARE OBVIOUSLY THE ANTIBIOTICS THAT ARE TESTED. THE RESISTANCE PHENOTYPE ALONG WITH SOME MEASUREMENT DATA, SO WE NOT ONLY ARE GOING TO GET IS IT RESISTANTS OR SENSITIVE BUT NUMERICAL DATA AS WELL AS THE TYPE THE WAY WE HAVE BEEN DOING TYPING, THE WAY WE TYPE THESE THINGS, ALONG WITH THE TESTING STANDARD. EUROPE USES A DIFFERENT STANDARD, WE WANT TO MAKE SURE THINGS ARE CALLED RESISTANT OR SENSITIVE PEOPLE KNOW HOW THAT DETERMINATION IS MADE. VERY BRIEFLY DESCRIBE PATHOGEN DETECTION PIPELINE FOR PURPOSES OF THIS TALK, WE DEVELOPED NCBI THIS BEFORE I CAME STARTED BEFORE I CAME TO NCBI, THE ABILITY TO TAKE DATA FROM THE SHORT READ ARCHIVE, THIS IS RAW SEQUENCING DATA. BUILD GENOME ASSEMBLIES WITH THAT DATA USING TWO STAGE TWO PHASE PROCESS, WE HAVE AA CRUDE ASSEMBLY, COMBINED ASSEMBLY, WE ANNOTATE, IDENTIFY THE GENES ON THAT GENOME. AND THEN WE CAN TURN THAT INTO GLOBAL -- WHAT WE WORKED ON NOW IS TURNING THAT INTO A REPORT SENSE TO RESEARCHER OR CLINICIAN, GIVING THAT A FIVE MEGA BASE FILE ATCG IS NOT GOING TO HELP THEM. SO HOW DO WE GO FROM THIS SET OF PROTEINS AFTER ANNOTATION, WHAT DO WE DO THEN? WE BUILT THE DATABASE WITH OVER 3600 RESISTANCE PROTEINS. DATABASE THEN WE TAKE THE SET OF PROTEINS WE SUBJECT IT TO HIDDEN MARK OFF MODELS IN BLACK TO IDENTIFY RESISTANCE GENES IN THE GENOME AND WE GENERATE THE REPORT. BRIEFLY DESCRIBE THE DATABASE, THIS DATABASE IS CONSTRUCTED WITH MULTIPLE SOURCES OF INPUT. WE HAVE DOMAIN EXPERTS WHO ARE PEOPLE WHO LITERALLY PUBLISH ON THIS IS WHAT YOU SHOULD CALL THIS GENE, SO THEY HAVE BEEN VERY HELPFUL. WE INCORPORATED DATA FROM SEVERAL GROUPS, THAT HAVE LARGE DATABASES, THE FDA CENTER FOR VETERINARY MEDICINE HAS BEEN A KEY PARTNER. REDS FINDER WHICH IS A DANISH GOVERNMENT INITIATIVE OUT OF THE STAT SERUM INSTITUTE AS WELL AS THE CARD WHICH IS THE CANADIAN INITIATIVE ON ANTI-MICROBIAL RESISTANCE. IN ORDER TO RESOLVE CONFLICT WES ALSO USE THE LITERATURE. SO ONE THING WE HAVE DONE IS WE HAVE DEVELOPED A STANDARDIZED DESCRIPTIVE NOMENCLATURE FOR THESE RESISTANCE GENES. THIS IS IMPORTANT, THE FIRST STEP FIGURING OUT HOW TO GO FROM GENOTYPE TO RESISTANCE PHENOTYPE IS FIGURING OUT WHAT THE PROTEINS MIGHT DO. JUST TO GIVE YOU AN IDEA OF WHY THIS IS IMPORTANT, THERE ARE MULTIPLE COPIES OF THIS GENE FOUND IN GEN BANK OFTEN CALLED (INAUDIBLE) 23. IT PROBABLY DOESN'T MEAN ANYTHING TO YOU NOR EXPECT IT TO. I WOULD BE SOMEWHAT SURPRISED IF IT DID. OTHER GROUPS HAVE SINCE THE DATA CLASS D LACTIMASE. IT'S A GENE THAT'S ACTIVE AGAINST ANTIBIOTICS. THAT'S BETTER. BUT WHAT WE WANT TO KNOW IS THIS IS A HYDROLYZING CLASS D LACTIMASE, IT SOUNDS LIKE A MULTI-AND IT IS, HYDROLYZEING LACTIMAS MS. O'CONNELL: PERS EVERY EVENT. WE NEED TO BE ABLE TO ASSIGN THIS FUNCTION. WE ALSO WANT TO STANDARDIZE THE OUTPUT SO THAT WE HAVE -- SO OTHER GROUPS CAN EASILY USE THE DATA WITHOUT HAVING TO DO CLEANING UP OF THE DATA THEMSELVES. ONE THING THAT THIS IS JUST SOMETHING SOMEBODY ONCE ASKED ME, WE BUILT A SLIDE, WHY DO WE USE HIDDEN MODEL? NCBI WAS THE HOME OF BLAST, SOMEBODY NCBI DEVELOPS, WHY WE USE TWO TECHNIQUES BLAST AND MARK UP MODELS. WHAT I'M SHOWING HERE AND THERE ARE TWO REASONS, ONE IS THAT THESE BLAST ONLY APPROACHES HAVE TROUBLE ASSIGNING FUNCTION TO NOVEL WHEELS. ONE THING WE DISCOVER IS AS WE SEQUENCE MORE GENOMES WE'RE FINDING GENES THAT DON'T LOOK LIKE THINGS WE HAVE SEEN BEFORE. IF YOU SAY I NEED THE FIND IDENTICAL MATCH YOU ARE GOING TO RUN INTO PROBLEMS. THE OTHER THING IS THAT THE HIDDEN MARK UP MODEL AZOLLAS TO PLACE LAST PROTEINS IN FUNCTIONALLY DEFINED FAMILIES ON EXISTING DATA MODELS NOT BASED ON SOME ARBITRARY CUT OFFS. A TOY EXAMPLE OF THIS, WE MIGHT FIND A PROTEIN 100% IDENTITY TO A KNOWN GENE, AND CONFERS RESISTANCE TO CARBO PEN TIN AND OTHER BETA LACKTIMS. IT'S CLOSE ENOUGH WE CAN SAY PART OF THE KBC FAMILY LIKELY RESISTANT TO THESE GENES THAT'S USEFUL INFORMATION FOR MICROBIOLOGISTS. IDEALLY SOMEBODY WITH THAT GO TO CHARACTERIZE THIS BUT WE HAVE A VERY GOOD HINT OF WHAT IT DOES. WE MAY NOT ASSIGN IT TO AMLY BUT WE CAN SAY THIS IS A BETA LACTIMACE AND THIS COULD CONVINCE SOMEBODY TO TRY TO FIGURE OUT WHAT THIS THING DOES, AND WE CAN ALSO SAY THIS IS NOT A BETA LACTIMACE. WHILE THAT SOUNDS TRIVIAL WE FOUND THE MAJORITY OF THINGS THAT HAVE BEEN CALLED BETA LACTIMACE ARE NOT BETA LACTIMACES. SO -- AND WE'RE WORK ON CLEANING THAT UP IN THE VERSION WHICH WE HAVE CONTROL. >> SO HAY MOVE TO PREDICTING ANTIBIOTIC SUSCEPTIBILITY. ONE POINT TO NOTE IS THAT VARIOUS GROUPS CLINICAL GROUPS AND AGENCIES START TO USE SEQUENCE, TO PREDICT PHENOTYPE, SO RIGHT NOW WORK IS DONE PHENOTYPICICALLY USING BASICALLY HIGH THROUGH PUT METHODS ORIGINALLY DESIGNED HALF A CENTURY AGO. AND HAVE BEEN SINCE REFINED BUT FOR INSTANCE THE PROGRAM THREE AGENCY CONSORTIUM THEY BASICALLY HAVE -- THEY'RE GOING TO DO LITTLE PHENOTYPES AND MOST GOING TO DETERMINE WHAT THEY'RE FOOD BORN PATHOGENS ARE RESISTANCE BY LOOKING AT GENOMES. THIS IS A REALITY. I'M SHOWING YOU DOWN HERE PAPER FOR VETERINARY MEDICINE WHERE THEY LOOK AT 76 ECOLI GENOMIC DATA THEY HAVE HIGH SPECIFICITY AND SENSITIVITY SO WE'RE TOING A GOOD JOB DOING BASED ON PHENOTYPES. I SHOULD POINT OUT WE WORK WITH FDA CENTER FOREVER VETERINARY MEDICINE AND ABLE TO IMPROVE THEIR GENETYPIC ASSESSMENT IN 7 OF 76 STRAINS SO THE WORK WE HAVE DONE IS MADE -- THEY ARE EAGER WORKING TO WORK WITH US TO TRY THE MAKE THEIR PROCESSES EVEN BETTER AND TO USE SOME OF OUR THINGS WE'RE DEVELOPING. SO JUST VERY BRIEFLY TALK ABOUT THE EMR REPORT, I WON'T GO THROUGH THIS IN EXCRUCIATING DETAIL. OBVIOUSLY WE TOLD WHAT GENES THEY HAVE AND WHAT FUNCTIONS BUT WE ALSO GIVE A LOT OF INFORMATION WHERE THEY OCCUR, THEIR GENOMIC CONTEXT, THIS IS REALLY IMPORTANT BECAUSE IT QUESTIONS SUCH AS DOES THIS PARTICULAR RESISTANCE GENE, IS IT ABLE TO BE TRANSFERRED WITH THESE TWO OTHER RESISTANCE GENES. THAT'S ACTUALLY A VERY IMPORTANT QUESTION. THE CLINICAL GROUPS WE WORK WITH KEEP SAYING WE NEED TO KNOW THIS.w3 KAREN FRANK GAVE A TALK WHERE SHE SAYS I DON'T WORRY ABOUT CHROMESOMALLY RESISTANT. I WORRY ABOUT MOBILE ELEMENTS MOVING FROM ONE BACKGROUND TO ANOTHER THAT'S HIGHER LEVEL OF ALERT SO KNOWING THE POSITION TO HAVE GENES IS CRITICAL. ONE THING WE WANT TO KNOW IS DO MR GENES TRAVEL IN TANDEM. I HAVE BLOWN UP SO YOU CAN READ IT, PART OF THE REPORT, WHAT I HAVE HIGHLIGHTED HERE IN RED AND THE GREEN ARE TWO CASES WHERE WE CAN EASILY IDENTIFY RECESS STANCE GENES THAT ARE CO-OCCURRING IN THE SAME SEGMENT OF THE GENOME. THESE THREE HERE WE'RE CERTAIN OCCUR ON ONE MOBILE PLASMID. SO FOR INSTANCE, SELECTION FOR CARBO PEN TIN RESISTANCE CONFERS RESISTANCE TO PHOSPHOMYCIN AS WELL AS OTHER DRUGS. THE OTHER THING WE CAN DO I ALLUDED BEFORE WE STARTED TO IDENTIFY WORKING WITH OUR CLINICAL COLLABORATORS NOVEL RESISTANCE GENES AND THE SHV FAMILY THAT HAS OVER 300 CURRENTLY IDENTIFIED DIFFERENT ALLELES, WE HAVE FOUND LOOKING AT A HANDFUL OF STRAINS NEW THINGS. SO WE'RE ACTUALLY GOING OUT THERE AND DISCOVERING NEW DIVERSITY THAT NEED TO BE CHARACTERIZED. SO BRIEFLY, WE SPENT A LOT OF TIME TRYING TO MAKE THIS WORK, BECAUSE ULTIMATELY ONE OF THE REASONS WE WANT TO WORK WITH CLINICAL GROUPS, THIS TIES INTO THE DISCUSSION HEARD EARLIER THIS MORNING WE NEED TO CONVINCE RAW SEQUENCE DATA. IT'S LARGE, BACTERIAL GENOMES ABOUT THE FOURTH OF A SIZE OF NETWORKS. WE NEED A VALUE-ADDED BACK IF WE CAN, W IF THEY WANT IT. WE HAVE BEEN TRYING, WE HAVE TWO DATA SETS, KNOWN TRUTH, THESE ARE GENOMES THAT HAVE BEEN CHARACTERIZED, THEY ARE WHAT WE CALL FINISH GENOMES, WE KNOW EVERY KNEW THEO TIDE IN THAT GENOME, WE KNOW THE ANSWER. THIS IS OUR POSITIVE AND NEGATIVE CONTROL FOR SEEING HOW WELL WE HAVE DONE. I WILL FOCUS ON TEN ISOLATES FROM THE CARBO PEN TIN RESISTANT OUTBREAK AT THE NIH HOSPITAL A COUPLE OF YEARS AGO. THE OTHER TYPE OF DATA WE HAVE, THIS IS GOING TO REFLECT WHAT WE WILL BE GETTING. GET SEQUENCE DATA, IT'S NOT THE TOP OF THE LINE WE CAN SOLVE ALL THE PROBLEMS TYPE OF DATA. WE HAVE IT FOR A BUNCH OF STRAINS ANTIOBIOGRAMS AS WELL AS THIS DATA. THIS IS GOING TO REFLECT WHAT THE KIND OF WHAT WE GET AN DAILY BASIS. WHAT WE DID, IF YOU WORK ON ANTI-BUY MICROBIAL RESTAINS, IT'S THE MOST IMPORTANT THING ARGUABLY, ONE OF THE MOST IMPORTANT THINGS IN MICROBIOLOGY. FROM GENOMICS STAND POINT IT IS A NIGHTMARE TRYING TO ASSEMBLE THIS PART OF THE GENOME. THIS IS JUST A SKETCH TO SHOW YOU WHY THIS IS A STRAIN WITH RESISTANT TO EVERYTHING WE CAN THROW AT IT. I BELIEVE IT WAS FOUND IN BRIGHAM AND WOMEN'S HOSPITAL A COUPLE OF YEARS AGO. HERE IS THE LIST OF GENES THAT SHOW UP, RESISTANCE GENES. HIGH HEIGHTED IN RED THE THINGS THAT ARE GOING TO GIVE US PROBLEM. HERE IS A CASE, EACH OF THESE COLUMNS, CHROMOSOME, IF IT OCCURS ON ONE OF THREE DIFFERENT PLASMIDS, WHAT YOU CAN START TO SEE IS MULTIPLE COPIES OF THIS IN DIFFERENT LOCATIONS SO WE'RE INTERESTED ABOUT THAT LINKAGE QUESTION, IT'S GOING TO BE VERY HARD TO ASSEMBLE. WE HAVE CASES WHERE WE HAVE TWO GENES THAT ARE ONE NUCLEOTIDE APART AND THAT TOO PRESENTS A CHALLENGE OCCURING IN MULTIPLE PARTS OF THE GENOME AS WELL AS WE HAVE THIS PARTICULAR CASE WHERE WE HAVE THREE COPIES OF A SINGLE GENE OCCURRING ON ONE PLASMID AND A FOURTH ON THE OTHER. FOR THOSE WHO DON'T KNOW GENOME ASSEMBLY, THIS IS DOING A JIG SAW PUZZLE WHERE YOU HAVE MULTIPLE IDENTICAL PIECES AND SOMEBODY LIT THE BOX COVER ON FIRE BECAUSE WE DON'T KNOW THE ANSWER WE'RE LOOKING FOR. SO HOW DID WE DO? OUR FALSE POSITIVES AND WHAT FASTTIVES ARE, WE SAID THERE'S A RESISTANCE GENE AND IN FACT THERE ISN'T ONE, WE HAVE 14 CASES OF THAT AND TWO OUT OF TEN SAMPLES. WHAT THIS WAS IS CONTAMINATION, I WILL TALK ABOUT HOW WE'RE GETTING AROUND THAT AND IN A SECOND. SOME GIVE US GOOD DATA, SOME GROUPS DON'T GIVE US GOOD DATA SO CONTAMINATION BETWEEN DIFFERENT BACTERIAL SAMPLES IS PROBLEM. WE ALSO HAD FALSE NEGATIVES EIGHT OUT OF 100 MAR GENES AN SIX OUT OF TEN SAMPLES WE FIGURED HOW TO CORRECT THIS BUT MOST OF THESE WERE ASSEMBLY PROBLEMS. SO HOW WELL ARE WE ASSEMBLING, THIS IS SOMETIMES JUST A LIMITATION OF TECHNOLOGY. COMPARED TO TO WHERE WE WERE WHEN I STARTED GENOMICS THIS IS AMAZING BUT NOT PERFECT. ONE KEY LIMITATIONS WE RUN INTO AS ALLUDED IN THE PREVIOUS SLIDE MULTIPLE COPY OF RESISTANCE GENES ARE HARD TO PUZZLE OUT ESPECIALLY TRYING TO SAY RESISTANCE GENE LINKED TO RESISTANCE GENE B. SO HOW -- THE FIRST ISSUE ARE THESE FALSE NEGATIVES. THIS IS REALLY BAD. IMAGINE IF YOU TOLD A CLINICIAN NO EWER HIGHLY RESISTANT ISOLATE DOESN'T HAVE A CARBO PEN TIN, OOPS WE GOT THAT WRONG THAT'S A BAD THING TO DO. WE DO NOT WANT FALSE CLAIMS OF SUSCEPTIBILITY. SO WHAT WE HAVE DONE IS DEVELOPED WHAT WE CALL TARGETED DE NOVO ASSEMBLY, IT'S UNDER DEVELOPMENT BUT IT SEEMS TO BE WORKING. WHERE WE USE THE ANTI-MICROBIAL RESISTANCE DATABASE, THOSE SEQUENCES TO PULL OUT THE READS THE RAW SEQUENCE DATA THAT CORRESPONDS TO ANTI-MICROBIAL RESISTANCE GENES. WE ASSEMBLED THOSE INTO RESISTANCE GENE SEQUENCES AND WE IDENTIFIED THE GENES FROM THAT PILE OF SEQUENCES. THERE ARE LIMITATIONS WE CAN ONLY ASSEMBLE AMR GENES THAT ARE SIMILAR TO KNOWN GENES. IF TOO DIVERGENT WE'LL NEVER P P PICK UP THE REED SO WE SEE THIS AS A COMPLIMENTARY APPROACH TO ASSEMBLIES THAT WE USE. JUST TO GIVE YOU AN EXAMPLE OF WHERE THIS MATTERED, HERE IS A LIST OF GENES FROM INCREDIBLY RESISTANT ORGANISM. THIS IS THE KNOWN TRUTH BECAUSE WE HAVE THE WHOLE GENOME SEQUENCE. WE DID A VERY GOODS JOB, WE BASICALLY CAPTURE EVERYTHING. EXCEPT WE WOULD HAVE CONCLUDED THAT THIS STRAIN IS SENSITIVE TO SHULFANAMIDE. WITH THE BETHE TA READS APPROACH WE DIDN'T GET ALL FOUR BUT WE WERE ABLE TO SAY THERE'S SULFAN,AMIDE RECESS STANCE GENES IN THESE ORGANISMS. IF WE TAKE THE UNION WE GET A GOOD ESTIMATE OF THE RESISTANCE GENES WITHIN A GENOME. THE OTHER THING I TALKED ABOUT IS FALSE POSITIVES. I MENTION THIS WAS CONTAMINATION. WHAT I'M SHOWING YOU HERE, HERE RESISTANCE PHENOTHAT WE WOULD CONCLUDES GENES WE SUPPOSE WE FOUND. BUT NOTED IN BLACK THESE ARE THINGS THAT ARE TRUE POSITIVES, WE KNOW THEY'RE IN THAT GENOME. AND THESE ARE THINGS -- THESE ARE GENE WEASANDS ARE IN THAT GENOME BUT IN FACT BASED ON THE FINISH GENOME SEQUENCE THE KNOWN ANSWER ARE NOT THERE. ONE THING TO NOTE WE WOULD HAVE CONCLUDED THE STRAIN IS RHESUS STAN TO GLYCOSIDES AND SHULF,NAMIDES AND THAT'S THE WRONG ANSWER. THE REASON WE KNOW THAT'S WRONG IS BECAUSE MOST OF THESE GENES WE SEQUENCE TO EXTREMELY HIGH DEPTH IN EACH OF THESE GENOMES. ESSENTIALLY YOU CAN THINK OF IT AS 300 -- THIS GENE IS SEQUENCED ON AVERAGE 300 TIMES. THAT'S HOW MUCH DATA WE HAVE. FOR THESE THINGS THAT ARE CONTAMINATION THEY ONLY SHOW UP TWO OR THREE OR SIX TIMES. SO THESE ARE CLEARLY FALSE POSITIVES WE THINK THEY'RE CONTAMINATION FROM SOURCE OR DURING SEQUENCE ACROSS. SO WE INSTITUTED FILTERS TO RULE. JUST ONE MORE THING ABOUT THIS, WHEN WE LOOK AT REAL LIFE DATA WE DON'T HAVE A KNOWN ANSWER. WE DID FAIRLY WELL WITH EXTENDED SPECTRUM BETA LACTIMACE GENOME, 127 OUT OF 139 TIMES, WE GOT ANSWER RESISTANT WITH THE RESISTANCE PHENOTYPE. WE HAD TEN PROBLEMATIC GENOMES, ASSEMBLY GENOMES WE DIDN'T GET THE WHOLE GENE BUT FIGURED SINCE I PUT THE SLIDE POT WHICH IS LAST WEEK. WE HAVE FIGURED OUT A TRIP AROUND THIS. SO WE CAN NOW GET RID OF THIS THESE, WE HAD TWO WITH FALSE POSITIVE, THAT ISSUE OF THIS OTHER DNA SNEAKING IN. WE FOUND TWO CASES OF CARBO PENT MACE, WE LOOKED AT 93 GENOME CARBO PENTIM RESISTANT GENOMES. AND THIS TOO WAS CONSISTENT WITH THE ANTS BIOGRAMS WE LEARNED SOME THINGS IN DOING THAT, AS WELL. AND WE ALSO DID THE DEPARTMENT OF DEFENSE PROVIDED US WITH SIX GENOMES THAT WERE A WHITMAN SAMPLER OF SPECIES AND WE GOT THE CORRECT ANSWER THERE. SO WE'RE DOING WELL. THERE ARE A BUNCH OF POTENTIAL COMPLICATIONS, I WON'T GO THROUGH IN DETAIL BUT THESE ARE THINGS WE'LL WORK ON IN THE FUTURE. I FEEL -- WANT TO FEAR OPTIMISTIC BECAUSE WE HAVE DONE A GOOD JOB SO FAR IN TERMS OF ADDICTION AND HIGHLIGHTS MORIA BOTH GRAMS IN AS WELL AS MORE DATA SO WE CAN GET A HANDLE ON THIS. FINALLY ONE OTHER THING WE'RE DOING, THIS IS -- I GAVE THIS TALK AT ASM MEETING I HAVE TO REMIND PEOPLE I'M TECHNICALLY PART OF NATIONAL LIBRARY OF MED SEN AND CONSEQUENTLY LIBRARIAN SO CURE RATION IS SOMETHING THAT I DO. AND THIS IS SOMETHING WE'RE DOING, THESE ARE -- WE'RE CURE RATING THE BETA LACTAM PROTEINS, THESE ARE A CRITICAL HEALTH CONCERN AS I MENTIONED EARLIER, PARTICULARLY CARBO PENTIM RECESS STANCE, IF A STRAIN IS RESISTANT TO THAT IT'S RESISTANT TO ALL OTHER BETA LACTIMS. REAL WORLD THOSE ARE RESISTANT TO EVERYTHING ELSE. A CRITICAL HEALTH PROBLEM IDENTIFIED BY THE WHITE HOUSE. PROTEIN ALLELE CURE RATION OF THESE GENES GIVES US A COMMON LANGUAGE SO IF I'M EPIDEMIOLOGIST AND I SAY I HAVE THIS OX 23 MORE IMPORTANTLY I HAVE SOMETHING THAT LOOKS ALMOST LIKE OX 23, HOW WOULD YOU LIKE TO FIGURE OUT WHAT IT DOES, THAT GIVES US A COMMON LANGUAGE, WE NEED THAT. IT'S A GOOD SYSTEM BECAUSE IT WAS PROBABLY CURATED BY ONE PERSON. AND THE PROBLEM IS HE'S RETIRED WE DID NOT WANT TO DISINTEGRATE IT'S TOO IMPORTANT TO THE FIELD. THE JOURNAL REQUIRES YOU GET A NAME TO PUBLISH ON THESE GENES. WE WERE ASKED TO TAKE THIS OVER, WE HAVE. IN COLLABORATION WITH PEOPLE THAT RAN THE DATABASE, WE CAME UP WITH REQUIREMENTS TO GET THE PROTEIN DNA SEQUENCE. WE ALSO STARTED REQUESTING THAT PEOPLE TAKE -- CLONE THE GENE INTO A COMMON BACKGROUND AND CHARACTERIZE WHAT THAT PROTEIN DOES IN THE COMMON BACKGROUND. THAT ALLOWS US TO SAY THIS PARTICULAR ALLELE CONFERS RESISTANCE, THESE ANTIBIOTICS, NOT OTHERS, USEFUL NOT JUST TO US BUT THE ENTIRE FIELD AS A WHOLE MAKING PREDICTIONS OR FIGURING WHAT A GENE DOES. AND THIS REALLY IS A CRITICAL THING WE'RE DOING BECAUSE MANY CASES A SINGLE AMINO ACID PROTEIN CAN CONFER RESISTANCE TO A WHOLE SLEW OF ANTIBIOTIC. THESE ARE FOR INSTANCE THE OXIN FAMILY ONE AMINO ACID TURNS TO A CARBO PENTIMASE. YOU POINT OUT WE OTHER ALSO HOSTING A TELECONFERENCE DOMAIN EXPERTS TO IMPROVE THE REQUIREMENTS IN THIS AREA, DO WE SEE DATA FOR ALL THESE GENES. THAT KIND OF THING. QUICKLY DESCRIBE FUTURE DIRECTIONS ONE THING WE HAVEN'T TACKLED YET IS ALLELIC VARIATIONS SO GENES THAT EVERY E. COLI HAS, VARIATION CONFERS RESISTANCE TO PARTICULAR ANTIBIOTICS. WE'RE STARTING TO LOOK AT THAT. WE ALSO WANT TO BE ABLE TO WORK ON FB. TB IS A BACTERIUM BUT MANY WAYS BUT DIFFERENT FROM OTHER TYPES OF ANTIBIOTIC RESISTANCE. GENOME WIDE ASSOCIATION STUDIES IN SOME RESPECTS BECAUSE IT'S ENTIRELY POINT MUTATIONS. YOU WANT TO TRY TO COLLECT MORE INFORMATION WHAT A SPECIFIC ANTIBIOTIC RESISTANCE GENE DOES, WHAT ANTIBIOTICS DOES THAT GENE CONFER RESISTANCE TO. ONE IS SOME OF THESE STUDIES ARE 20-YEAR-OLD OLD AND SOME THAT EXIST TODAY DIDN'T EXIST BACK THEN. WE ARE WORKING BRIGHAM AND WOMEN'S HOSPITAL AS WELL AS NIAID FUNDED GENOME SEQUENCING CENTERS TO IMPROVE THE SUBMISSION PROCESS. WE KNOW TO A CERTAIN EXTENT THIS IS VERY IMPORTANT. WE DON'T WANT TO LOSE THE INFORMATION, PARTICULARLY HOSPITALS ARE PRODUCING. THEY ARE NOT UNDER OBLIGATION LIKE FEDERALLY FUNDED GENOME RESEARCHERS ARE TO SUBMIT THEIR DATA. WE WANT TO GET THAT DATA SWOONED TO MAKE THAT DATA USEFUL TO THEM AND EASY TO SUBMIT. FINALLY, WE'RE ALSO WORKING ON THIS LOCK TERM ISSUE OF RESOLVING NOMENCLATURE NOT JUST BETA LACTIMACE BUT SERIES OF GENES ANTIBIOTIC RESISTANCE GENES. AND IT'S WORSE, AT LEAST GEORGE HE WAS COMMITTEE OF ONE, REALLY ARE NO STANDING COMMITTEES. SO IT'S TOWER OF BABEL IN MANY WAYS. JUST THE PEOPLE CREDIT THAT'S DUE, THERE'S A WHOLE GROUP OF PEOPLE WHO DEVELOPED A PATHOGEN DETECTION PIPELINE, THE SHOULDERS THIS EMR PROJECT STANDS ON. THERE ARE THREE WHO WORKED EXTENSIVELY ANTI-MICROBIAL RESISTANCE COMPOUNDS, (INDISCERNIBLE) HEADS UP THE ENTIRE PATHOGEN DETECTION PROJECT AND JIM OSTEL INTENSIVELY INVOLVED IN ADVISING THIS PROJECT. AND ALSO I WANT TO ACKNOWLEDGE COLLABORATORS, AS WELL. I'LL STOP THERE. FOR A DISCUSSION. [APPLAUSE] >> THANK YOU, (INAUDIBLE) -- THE LAB DEAL WITH ON AN INDIVIDUAL BASIS GETTING FOLKS READY IN PUBLIC HEALTH TO BE ABLE TO DEAL WITH THE PROBLEM OF ANTI-MICROBIAL RESISTANCE, GENOME SEQUENCING OF MANY, MANY PATHOGENS. I THINK THE COMMENT -- TWO COMMENTS THE MAKE, ONE THE BIGGER ISSUE IS MANY PEOPLE WORKING ON SEQUENCING IN MANY INSTITUTIONS AND THE TOWER OF BABEL WITH THE KNOWLEDGE WE NEED ORGANIZATION NOT TO TAKE IT OVER BUT TO AUTOMATE IT TO BRING THE GROUPS TOGETHER. WE DON'T TALK THE SAME LANGUAGE. THE OTHER THING IS THAT THIS IS PREPARATORY WORK THE ALGORITHM THAT I WAS TALKING ABOUT YESTERDAY, WHERE THERE WILL BE TOOL SUPPOSITION TO SAY WITH THIS GENOME THAT THIS PATIENT DRUG HISTORY, THESE ARE THE BEST DRUGS THAT YOUR PATIENT IS LIKELY TO BENEFIT FROM. THIS IS THE INCREDIBLY IMPORTANT PREPARATORY WORK TO GET TO THAT POINT. >> I THINK THIS WORK IS REALLY EXCITING. AND IN TERMS OF WHAT WE HAVE BEEN TALKING ABOUT TEACHING PLANNING WHERE WE'RE GOING, ATTEMPTING TO DEAL WITH GREATER STANDARDIZATION AND TERMINOLOGY AND COORDINATING OF THINGS, THAT IS A VERY TO MY WAY OF THINKING THAT SOMETHING NEEDS TO BE CONSIDERED BY THE BOARD HOW MUCH AN INVESTMENT WE WILL MAKE IN THAT BECAUSE THAT TYPE OF WORK IT SEEMS TO ME LINES UP VERY WELL WITH EXPERIENCE AND OBVIOUSLY THE ABILITY TO DISSEMINATE AND MAKE READILY AVAILABLE STANDARD NOMENCLATURE AN HELP THEM USING IT SO I GENERALLY FEEL SINCE I KNEW THERE WAS SUCH A THING AS GENE AND GENE NOMENCLATURE I HEARD PEOPLE SAY THE (INAUDIBLE) IN THAT AIR IS DISASTER -- THAT AREA IS DISASTER. THE MOST CONFUSING THING IN THE WORLD. SO HAVING BEEN INVOLVED IN THE UMS PROJECT IN THE BEGINNING I UNDERSTAND THAT THAT WAS SIMPLE IN COMPARISON AND I KNOW THAT IT ACTUALLY WASN'T SIMPLE, BACK WHEN WE DID IT. BUT THIS DOES SEEM TO BE THE AREA THAT DESERVES REAL CONSIDERATION IN TERMS OF A GREATER INVESTMENT INVOLVED BY NLM. AND I AGREE WITH THE NOTION OF CARD COORDINATION OF GROUPS SO FORTH. WE WERE PROBABLY HAVING EXPANSION OF STAFF WHICH MIGHT NOT BE REQUIRED IF WE THOUGHT WE WERE GOING TO DO THIS OURSELVES AND WOULDN'T HAVE ALL THE EXPERTISE TO DO IT ANYWAY. >> I THINK THE ISSUE TO WORRY ABOUT HOE THE (INAUDIBLE) SEQUENCING, HAPPENING IN MANY UNIVERSITIES AND I DON'T WANT TOWS GET TO A POINT WHERE WE HAVE A WHOLE MESS TO DEAL WITH, IT'S REALLY SIGNIFICANT. >> I HEAR YOU. >> TWO QUESTIONS, WHEN DO YOU THINK IT WILL BE TO THE POINT THAT CAN TAKE A SAMPLE OF CSF OR BLOOD FROM RESUSCITATION AND PUT ANYTIME A MACHINE AND IN FIVE MINUTES NOTE THE BACTERIA AND THE FLOW THAT IS MOST USEFUL TO USE? THERE IS A BIT OF THAT SO FIVE MINUTES, I DON'T KNOW. BUT I CAN TELL YOU THAT THERE ARE ALREADY GROUPS THAT ARE LOOKING AT RNA AND THEY'RE ABLE TO DO IT IN FOUR TO EIGHT HOURS FROM URINE. ANYTHING THAT HAS HUMAN CELLS IS A PROBLEM FOR TECHNICAL REASONS. WE DON'T WANT TO LOOK AT HUMAN, WE WANT THE BACTERIUM BUT THERE'S TOO MUCH DNA OFTEN. BUT IN TERMS OF THE RAPIDITY, IN TERMS OF BIOINFORMATICS SIDE IN PARTICULAR, WE ALREADY HAVE THE CAPACITY, I MEAN, WITHIN -- NOT JUST LOTS OF GROUPS HAVE THE CAPACITY IF THEY GET A SEQUENCE WITHIN HOURS TO BE ABLE TO SAY THIS IS THE COMPLIMENT OF RESISTANCE GENES, IT COULD GO FASTER. ARE GENERATING OUR REPORT FROM SET OF PROTEINS TAKES TEN SECONDS. IT'S A TRIVIAL THING. BUT I THINK THE REAL STUMBLING BLOCKS THERE ARE ON THE DNA EXTRACTS SIDE BEFORE THE BACTERIUM. BUT RIGHT NOW LIKE I SAID THERE ARE PEOPLE WHO I HAVE A COLLEAGUE WHO WORKS ON KIDS WITH NECROTIZING COLITIS AND THEY ASSEMBLE THE DOMINANT BACTERIUM IN THOSE GUTS WHICH IS THOUGHT TO BE THE ETIOLOGICAL AGENT IN SOME CASES, THEY'RE ABLE TO DO THAT RATHER QUICKLY. I THINK RIGHT NOW THE REAL ISSUE IS THAT THERE IS A GAP, I WORKED ON THIS WHEN I WAS AT ROAD TO START TO DEVELOP THIS, THERE IS A REAL GAP, IT HASN'T BEEN CARRIED TO MARKET SO THERE'S ACADEMIC GROUPS THAT CAN DO THIS QUICKLY. YOU CAN DO WHAT WE CALL AT BROAD A FIRE DRILL, WE DROP EVERYTHING AND DO THIS, TURN AROUND IN 12 TO 18 HOURS. BUT THAT IS NOT A SUSTAINABLE MODEL, WHAT TECHNOLOGICAL IMPROVEMENTS NEED TO HAPPEN. ON THE INFORMATICS SIDE WE'RE AHEAD OF THE GAME COMPARED TO THE DNA EXTRACTS SIDE. >> SO THERE'S NO COMPLICATION (INAUDIBLE) >> I DON'T WANT THE SAY THERE'S NO COMPUTATIONAL PROBLEMS BUT THEY PALE. WE'RE NOT THE RATE LIMITING STEP. >> ANOTHER USAGE OF THIS IS AGRICULTURAL PART (INAUDIBLE) CALL IT FORENSIC GENOMICS. >> I'M GLAD YOU BROUGHT THAT UP. ONE CUSTOMER OF THE DETECTION PATHOGEN PROJECT HEADED OUT OF FDA, AND THEY ARE SEQUENCING -- I THINK THEIR PLANS ARE TO SEQUENCE 8, 10,000 ISOLATE THIS ISIER ASSOCIATED WITH FOOD BORN CONTAMINATION. AND SO IT'S ABSOLUTELY, I MEAN FORENSICS, I THINK THAT REALLY WAS THE INITIAL IMPETUS FOR THE PATHOGEN DETECTION PIPELINE TO HELP OTHER GROUPS WITH FORENSIC GENOMICS, FORENSIC ANALYSIS. SINCE THEN WE GOT REQUESTS SAYING WE WANT TO KNOW ABOUT ANTI-MICROBIAL RESISTANCE. ABSOLUTELY. I THINK RELATED TO THE ISSUES I RAISED HERE, ONE QUESTION WE START WITH GENOME SEQUENCES PARTICULARLY GOOD ONES IS DO WE SEE LET'S SAY A PLASMA WITH THREE RESISTANCE GENES, ONE WE'RE CONCERNED ABOUT, TO WE SEE THAT IN AN AGRICULTURAL SETTING AND DO WE ALSO SEE THAT IN A HOSPITAL NEARBY. RATHER THAN SAYING PHENOTYPES OR A BAG OF GENES WE CAN DRILL DOWN SAY THAT'S THE EXACT SAME GENOMIC CONTEXT IN THIS STRAIN, THAT STRAIN, THAT'S IMPORTANT FOR TEASING APART, FIGURING OUT WHETHER -- WHAT THE LINKS ARE BETWEENING A DULL CHURL USE AND CLINICAL USE IN TERMS OF RESISTANCE. ABSOLUTELY. RIGHT NOW I WOULD SAYING CULTURAL IS OUR BIGGER CUSTOMER THAN CLINICAL USE. >> OTHER QUESTIONS OR COMMENTS? OKAY. THANK YOU VERY MUCH. [APPLAUSE] >> SO THAT'S ALL OF THE BUSINESS PART THAT WE HAVE. ARE THERE ANY OTHER ANNOUNCEMENTS OR THINGS THAT PEOPLE WANT TO SHARE? I KNOW THAT'S EITHER -- AT THE -- FOR THOSE HERE THIS AFTERNOON AND INTERESTED IN HEARING ABOUT THAT HISTORY OF NURSING INVOLVEMENT IN GETTING THE HEALTHCARE SYSTEM TO PAY MORE ATTENTION TO EVIDENCE OF VIOLENCE AGAINST WOMEN AND TO TRY TO DO THINGS TO HELP THEM. WE ARE ARE HAVING THE THE EXHIBITION IN THAT AREA THIS AFTERNOON AT ONE O'CLOCK. IT WILL BE IN THE LISTER HILL AUDITORIUM. FOR THOSE OF YOU WHO -- I THINK WE PASSED OUT THE INFORMATION YESTERDAY, YOU WANT TO HEAR MORE ABOUT THE PMI THAN YOU HAVE ALREADY HEARD SOME OF YOU I ARE GOING TO BE OVER HERE TELLING US, THAT THERE IS THE INFORMATION ABOUT THE WEBINAR AND DIAL-IN INFORMATION THAT WAS GIVEN TO YOU. IF IT'S NOT GOING TO BE WEBINAR WHERE THERE'S ALSO AN AUDITORIUM WITH -- WHERE YOU CAN GO SIT. I THINK IT'S ALL OVER THE -- SO IF YOU'RE NOT PRESENTING AND YOU'RE CALLING IN, IT'S NOT ACTUALLY AN AUDITORIUM TYPE EVENT. >> WILL THERE BE A PUBLISHED REPORT? >> ABSOLUTELY. I SPICK THAT >> GET THEIR FORMAL PRESENTATION FROM THEIR WORK FROM THE WORKING GROUP. AND THEN THEY WILL DISCUSS THAT AND REASONABLE EXPECTATION THEY WILL ACCEPT THE REPORT AND PASS IT ON TO FRANCIS. AND UNLESS THEY SAY MAKE SOME CHANGES TO THE REPORT, WHICH I DON'T ACTUALLY EXPECT OR ANYTHING SIGNIFICANT, I THINK THERE WILL BE A VERSION OF THE REPORT THAT COMES AVAILABLE VERY SOON. AND IT MAKES GOOD READING. >> I THINK THAT PART, JOSH, YOU DO THE TECHNICAL PART AND I'LL PUT IT IN SUMMARY FORM. >> IT MIGHT BE HARD TO READS -- (INAUDIBLE). (OVERLAPPING SPEAKERS) >> WELL, THANK YOU ALL VERY MUCH. SAFE TRAVELS, WE'LL SEE YOU NEXT YEAR. THANK YOU.