1
00:00:04,104 --> 00:00:06,073
BY DEVELOPING

2
00:00:06,073 --> 00:00:06,673
THE FIRST

3
00:00:06,673 --> 00:00:08,175
CELL FREE SYSTEM

4
00:00:08,175 --> 00:00:10,410
TO SUPPORT A VERTEBRATE CHROMOSOME.

5
00:00:10,410 --> 00:00:13,413
DNA REPLICATION USING FROG EGG EXTRACTS.

6
00:00:13,780 --> 00:00:15,015
AND THIS WORK TRANSFORMED

7
00:00:15,015 --> 00:00:17,618
THE STUDY OF DNA REPLICATION.

8
00:00:17,618 --> 00:00:20,687
IN 1999, DOCTOR WALTER JOINED HARVARD

9
00:00:20,687 --> 00:00:22,656
MEDICAL SCHOOL AS AN ASSISTANT PROFESSOR

10
00:00:22,656 --> 00:00:23,490
AND WAS LATER

11
00:00:23,490 --> 00:00:26,526
PROMOTED TO FULL PROFESSOR IN 2010.

12
00:00:26,526 --> 00:00:30,130
AND SINCE 2013, HE HAS ALSO BEEN

13
00:00:30,130 --> 00:00:31,632
AN INVESTIGATOR AT THE HARVARD

14
00:00:31,632 --> 00:00:33,200
HUGHES MEDICAL INSTITUTE

15
00:00:33,200 --> 00:00:34,768
AND DOCTOR WALTER REMAIN

16
00:00:34,768 --> 00:00:36,703
AS A DEDICATED EDUCATOR,

17
00:00:36,703 --> 00:00:38,105
TEACHING DNA REPLICATION

18
00:00:38,105 --> 00:00:39,473
AND REPAIR IN GRADUATE

19
00:00:39,473 --> 00:00:42,209
GRADUATE COURSES AT HARVARD.

20
00:00:42,209 --> 00:00:44,611
WE'RE THRILLED TO HAVE HIM HERE TODAY

21
00:00:44,611 --> 00:00:45,078
TO SHARE.

22
00:00:45,078 --> 00:00:47,047
HIS INSIGHTS AND EXPERTISE

23
00:00:47,047 --> 00:00:47,981
FOR THE AUDIENCE.

24
00:00:47,981 --> 00:00:48,482
SO PLEASE

25
00:00:48,482 --> 00:00:50,617
PUT ANY QUESTIONS IN THE CHAT BOX

26
00:00:50,617 --> 00:00:51,852
AND WE WILL GO THROUGH THEM

27
00:00:51,852 --> 00:00:53,420
AT THE END OF THE SEMINAR.

28
00:00:53,420 --> 00:00:55,422
AND PLEASE JOIN ME IN WELCOMING

29
00:00:55,422 --> 00:00:56,857
DOCTOR JOHANNES. WATER.

30
00:00:59,593 --> 00:01:00,827
WELL, THANK YOU VERY MUCH, CHUN,

31
00:01:00,827 --> 00:01:02,496
FOR THE VERY NICE INTRODUCTION.

32
00:01:02,496 --> 00:01:05,432
CAN EVERYONE HEAR ME?

33
00:01:05,432 --> 00:01:06,533
YES WE CAN.

34
00:01:06,533 --> 00:01:09,469
OKAY, GREAT.

35
00:01:09,469 --> 00:01:09,870
SO. YEAH.

36
00:01:09,870 --> 00:01:11,271
SO IT'S A REAL PLEASURE AND HONOR

37
00:01:11,271 --> 00:01:14,741
TO BE ABLE TO GIVE THIS TALK IN

38
00:01:15,142 --> 00:01:17,244
WHAT I GUESS IS THE ORIGINAL,

39
00:01:17,244 --> 00:01:20,080
GENOME MAINTENANCE WEBINAR.

40
00:01:20,080 --> 00:01:23,750
AND, I'M VERY EXCITED TO TELL YOU TODAY

41
00:01:23,750 --> 00:01:26,086
ABOUT HOW WE HAVE TACKLED

42
00:01:26,086 --> 00:01:27,587
A REALLY CLASSIC PROBLEM,

43
00:01:27,587 --> 00:01:29,957
THE PROBLEM OF HOW TRANSCRIPTION

44
00:01:29,957 --> 00:01:32,526
IS COUPLED TO REPAIR

45
00:01:32,526 --> 00:01:35,562
USING A FEW NEW APPROACHES.

46
00:01:35,963 --> 00:01:38,565
ONE OF THESE IS A CELL FREE SYSTEM

47
00:01:38,565 --> 00:01:41,568
THAT WE HAVE DEVELOPED THAT SUPPORTS TC-NER

48
00:01:42,035 --> 00:01:45,038
AND ANOTHER IS USING

49
00:01:45,205 --> 00:01:46,506
AN IN SILICO SCREENING

50
00:01:46,506 --> 00:01:46,907
METHOD

51
00:01:46,907 --> 00:01:48,108
TO LOOK FOR NEW PROTEIN

52
00:01:48,108 --> 00:01:50,610
PROTEIN INTERACTIONS.

53
00:01:50,610 --> 00:01:52,446
AND BY COMBINING THESE APPROACHES,

54
00:01:52,446 --> 00:01:53,613
WE THINK WE'VE MADE,

55
00:01:53,613 --> 00:01:56,450
AN IMPORTANT NEW OBSERVATION,

56
00:01:56,450 --> 00:01:58,185
WHICH IS REALLY,

57
00:01:58,185 --> 00:01:59,987
IDENTIFYING THE ROLE OF A PROTEIN CALLED

58
00:01:59,987 --> 00:02:02,022
STK19 AND COUPLING,

59
00:02:02,022 --> 00:02:04,691
TRANSCRIPTION STALLING TO THE DOWNSTREAM

60
00:02:04,691 --> 00:02:05,726
REPAIR EVENTS.

61
00:02:07,194 --> 00:02:07,828
SO I'M GOING TO

62
00:02:07,828 --> 00:02:10,831
BEGIN JUST WITH THE 30,000FT VIEW,

63
00:02:11,231 --> 00:02:13,934
WHICH IS TO SAY THAT COLLECTIVELY,

64
00:02:13,934 --> 00:02:15,335
AS A FIELD,

65
00:02:15,335 --> 00:02:16,837
WE ARE ALL TRYING TO UNDERSTAND,

66
00:02:16,837 --> 00:02:20,374
THE HIGHLY VARIED DNA REPAIR PATHWAYS

67
00:02:20,807 --> 00:02:25,278
THAT RESPOND TO AND FIX THE VERY DIVERSE

68
00:02:25,278 --> 00:02:27,547
LESIONS THAT ARE GENERATED BY EXOGENOUS

69
00:02:27,547 --> 00:02:30,217
AND ENDOGENOUS AGENTS.

70
00:02:30,217 --> 00:02:32,219
AND WE'RE ALSO, OF COURSE,

71
00:02:32,219 --> 00:02:33,420
TRYING TO UNDERSTAND

72
00:02:33,420 --> 00:02:35,956
HOW DEFECTS IN THESE PATHWAYS

73
00:02:35,956 --> 00:02:38,291
GIVE RISE TO THE VARIOUS DISEASES

74
00:02:38,291 --> 00:02:39,626
THAT WE SEE IN HUMANS,

75
00:02:39,626 --> 00:02:42,629
NOT ONLY, TO UNDERSTAND REALLY

76
00:02:42,629 --> 00:02:43,830
THE THE BASIS

77
00:02:43,830 --> 00:02:46,500
OF THESE PATHOLOGICAL STATES, BUT ALSO,

78
00:02:46,500 --> 00:02:48,101
IN SOME CASES TO ACTUALLY SEE

79
00:02:48,101 --> 00:02:49,469
IF WE CAN FIND

80
00:02:49,469 --> 00:02:52,873
SOME, SOME THERAPEUTIC INTERVENTIONS

81
00:02:53,206 --> 00:02:54,741
TO, TO AMELIORATE

82
00:02:54,741 --> 00:02:57,277
SOME OF THESE CONDITIONS.

83
00:02:57,277 --> 00:02:59,312
NOW, MY LABORATORY FOR MANY YEARS

84
00:02:59,312 --> 00:03:00,047
HAS FOCUSED

85
00:03:00,047 --> 00:03:01,681
ON REPLICATION

86
00:03:01,681 --> 00:03:03,950
COUPLED REPAIR OF VERY COMPLEX

87
00:03:03,950 --> 00:03:05,752
LESIONS LIKE DNA INTERSTRAND

88
00:03:05,752 --> 00:03:08,288
CROSS LINKS AND DNA PROTEIN CROSS LINKS.

89
00:03:08,288 --> 00:03:10,424
AND THAT INTEREST CONTINUES.

90
00:03:10,424 --> 00:03:11,691
BUT WHAT I'M GOING TO FOCUS ON

91
00:03:11,691 --> 00:03:14,361
TODAY IS A RELATIVELY NEW INTEREST

92
00:03:14,361 --> 00:03:15,629
IN THE MECHANISM

93
00:03:15,629 --> 00:03:16,997
OF NUCLEOTIDE

94
00:03:16,997 --> 00:03:19,599
EXCISION REPAIR, IN PARTICULAR,

95
00:03:19,599 --> 00:03:20,500
THE TRANSCRIPTION

96
00:03:20,500 --> 00:03:23,170
COUPLED BRANCH OF THIS PATHWAY.

97
00:03:23,170 --> 00:03:24,504
SO LET ME BEGIN

98
00:03:24,504 --> 00:03:25,639
BY GIVING YOU A LITTLE BIT OF

99
00:03:25,639 --> 00:03:27,274
BACKGROUND ABOUT WHAT WE KNOW

100
00:03:28,408 --> 00:03:30,110
ABOUT NUCLEOTIDE EXCISION REPAIR.

101
00:03:30,110 --> 00:03:32,746
THIS IS A PATHWAY THAT REALLY SPECIALIZES

102
00:03:32,746 --> 00:03:35,982
IN THE REPAIR OF VERY BULKY

103
00:03:35,982 --> 00:03:39,319
DNA LESIONS THAT DISTORT THE DNA DUPLEX.

104
00:03:39,319 --> 00:03:40,854
THESE CAN BE GENERATED

105
00:03:40,854 --> 00:03:43,623
BY UV LIGHT, CIGARETTE SMOKE,

106
00:03:43,623 --> 00:03:47,494
AND ALSO VARIOUS ENDOGENOUS ALDEHYDES

107
00:03:48,728 --> 00:03:50,597
AND THESE AGENTS

108
00:03:50,597 --> 00:03:51,965
LEAD TO THE FORMATION

109
00:03:51,965 --> 00:03:53,867
OF THESE BULKY ADDUCT

110
00:03:53,867 --> 00:03:55,502
THAT HAVE THE PROPERTY

111
00:03:55,502 --> 00:03:58,505
OF ACTUALLY DISTORTING THE DNA DUPLEX.

112
00:03:58,505 --> 00:04:01,007
AND IN ONE BRANCH OF NUCLEOTIDE EXCISION

113
00:04:01,007 --> 00:04:02,242
REPAIR, CALLED THE GLOBAL

114
00:04:02,242 --> 00:04:03,176
GENOMIC PATHWAY,

115
00:04:03,176 --> 00:04:03,877
WHICH CAN

116
00:04:03,877 --> 00:04:05,078
IN PRINCIPLE FUNCTION

117
00:04:05,078 --> 00:04:07,581
ANYWHERE IN THE GENOME.

118
00:04:07,581 --> 00:04:09,883
THE DISTORTION IN THE DNA DUPLEX

119
00:04:09,883 --> 00:04:12,853
IS RECOGNIZED BY A PROTEIN CALLED XPC

120
00:04:12,853 --> 00:04:14,988
XERODERMA PIGMENTOSUM

121
00:04:14,988 --> 00:04:16,456
GROUP

122
00:04:16,456 --> 00:04:17,324
C PROTEIN

123
00:04:17,324 --> 00:04:20,393
THAT IS, THAT IS MUTATED IN THIS,

124
00:04:20,393 --> 00:04:23,430
SKIN CANCER PREDISPOSITION SYNDROME.

125
00:04:23,430 --> 00:04:26,466
AND WHAT IT DOES IS REALLY QUITE CLEVER.

126
00:04:26,466 --> 00:04:30,370
IT ACTUALLY SENSES NOT THE LESION ITSELF,

127
00:04:30,370 --> 00:04:32,172
BUT THE DISTORTION IN THE DUPLEX.

128
00:04:32,172 --> 00:04:33,573
AND IT ACTUALLY GRABS

129
00:04:33,573 --> 00:04:36,042
HOLD OF THE NON-DAMAGED STRAND

130
00:04:36,977 --> 00:04:39,479
BY VIRTUE OF THE DESTABILIZATION

131
00:04:39,479 --> 00:04:40,413
OF THE DUPLEX

132
00:04:40,413 --> 00:04:43,416
THAT OCCURS AS A RESULT OF THE DAMAGE.

133
00:04:43,550 --> 00:04:47,187
AND XPC THEN FUNCTIONS TO RECRUIT

134
00:04:47,587 --> 00:04:50,590
THE GENERAL TRANSCRIPTION FACTOR TFIIH

135
00:04:51,091 --> 00:04:53,894
WHICH

136
00:04:53,894 --> 00:04:56,730
WHICH HAS A KEY ROLE

137
00:04:56,730 --> 00:04:59,199
IN INITIATING TRANSCRIPTION.

138
00:04:59,199 --> 00:05:00,934
SO IT ACTUALLY PROMOTES

139
00:05:00,934 --> 00:05:02,068
THE PROMOTER

140
00:05:02,068 --> 00:05:05,372
OPENING AT AT TRANSCRIPTION START SITES.

141
00:05:06,606 --> 00:05:09,142
AND IT HAS A PARTIALLY OVERLAPPING

142
00:05:09,142 --> 00:05:11,278
FUNCTION IN REPAIR.

143
00:05:11,278 --> 00:05:14,481
SO AFTER BEING RECRUITED BY XPC,

144
00:05:15,215 --> 00:05:18,852
THE XPB SUBUNIT OF TFIIH

145
00:05:19,219 --> 00:05:21,221
THIS IS A TRANSLOCASE.

146
00:05:21,221 --> 00:05:21,922
IT'S AN ATPASE

147
00:05:21,922 --> 00:05:24,558
IS THAT TRANSLOCATES ALONG DNA

148
00:05:24,558 --> 00:05:25,792
AND IT ACTUALLY PUMPS

149
00:05:25,792 --> 00:05:30,030
THE DUPLEX TOWARDS THE XPC PROTEIN.

150
00:05:30,030 --> 00:05:31,631
AND IN DOING SO,

151
00:05:31,631 --> 00:05:33,934
IT ROTATES AND UNDERWINDS

152
00:05:33,934 --> 00:05:35,168
THE DNA THAT LEADS

153
00:05:35,168 --> 00:05:37,370
TO LOCAL STRAND SEPARATION.

154
00:05:37,370 --> 00:05:38,738
AND NOW A SECOND

155
00:05:38,738 --> 00:05:41,741
SUBUNIT IN TFIIH, CALLED XPD,

156
00:05:41,908 --> 00:05:43,076
WHICH IS A FIVE PRIME

157
00:05:43,076 --> 00:05:44,578
TO THREE PRIME HELICASE,

158
00:05:45,612 --> 00:05:48,014
GRABS HOLD OF ONE STRAND

159
00:05:48,014 --> 00:05:49,282
AND THEN STARTS TRANS

160
00:05:49,282 --> 00:05:52,619
LOCATING TOWARDS THE SITE OF THE DAMAGE,

161
00:05:53,153 --> 00:05:55,388
AND IF DAMAGE IS PRESENT,

162
00:05:55,388 --> 00:05:58,225
THEN XPD GETS STUCK.

163
00:05:58,225 --> 00:06:01,595
AND THIS IS A PROCESS THAT IS CALLED

164
00:06:01,728 --> 00:06:02,996
LESION VERIFICATION

165
00:06:02,996 --> 00:06:04,698
OR DAMAGE VERIFICATION.

166
00:06:04,698 --> 00:06:07,167
AND THE STALLING OF XPD

167
00:06:07,167 --> 00:06:10,170
THEN NUCLEATES THE DOWNSTREAM EVENTS,

168
00:06:10,170 --> 00:06:11,104
WHICH INCLUDES

169
00:06:11,104 --> 00:06:13,006
THE RECRUITMENT OF STRUCTURE

170
00:06:13,006 --> 00:06:14,975
SPECIFIC ENDONUCLEASES XPF

171
00:06:14,975 --> 00:06:15,976
AND XPG

172
00:06:15,976 --> 00:06:19,212
THAT INCIZE THE DAMAGED STRAND AND,

173
00:06:19,412 --> 00:06:21,414
REMOVE A SHORT OLIGONUCLEOTIDE

174
00:06:21,414 --> 00:06:23,083
CONTAINING THE LESION.

175
00:06:23,083 --> 00:06:25,585
AND THEN THE FINAL EVENT

176
00:06:25,585 --> 00:06:28,722
IS A GAP FILLING STEP.

177
00:06:30,056 --> 00:06:32,259
SO THIS IS THE SO-CALLED GLOBAL

178
00:06:32,259 --> 00:06:33,159
GENOMIC PATHWAY,

179
00:06:33,159 --> 00:06:34,027
WHICH, AS I SAID,

180
00:06:34,027 --> 00:06:35,095
CAN FUNCTION,

181
00:06:35,095 --> 00:06:38,131
ESSENTIALLY ANYWHERE IN THE GENOME.

182
00:06:38,131 --> 00:06:39,032
BUT THERE'S A SECOND

183
00:06:39,032 --> 00:06:40,634
VERY IMPORTANT BRANCH

184
00:06:40,634 --> 00:06:41,868
OF NUCLEOTIDE EXCISION

185
00:06:41,868 --> 00:06:43,270
REPAIR CALLED TRANSCRIPTION

186
00:06:43,270 --> 00:06:44,070
COUPLED REPAIR.

187
00:06:44,070 --> 00:06:47,641
AND THIS COMES INTO PLAY WHEN THE GLOBAL

188
00:06:47,641 --> 00:06:50,644
GENOMIC PATHWAY FAILS TO REMOVE A LESION

189
00:06:50,844 --> 00:06:51,911
AND AN RNA

190
00:06:51,911 --> 00:06:54,914
POLYMERASE COLLIDES WITH THE DAMAGE.

191
00:06:55,015 --> 00:06:58,151
NOW, THIS WOULD BE A TRULY DIRE SITUATION

192
00:06:58,151 --> 00:06:59,452
BECAUSE,

193
00:06:59,452 --> 00:07:01,354
THE LESION STALLS THE POLYMERASE

194
00:07:01,354 --> 00:07:03,256
AND LOCALLY DISRUPTS,

195
00:07:03,256 --> 00:07:05,225
GENE EXPRESSION ON THE ONE HAND

196
00:07:05,225 --> 00:07:08,261
AND ON THE OTHER HAND, THE THE POLYMERASE

197
00:07:08,261 --> 00:07:09,763
ACTUALLY SHIELDS THE LESION

198
00:07:09,763 --> 00:07:11,564
FROM THE GLOBAL GENOMIC PATHWAY.

199
00:07:11,564 --> 00:07:12,232
SO THIS WOULD BE,

200
00:07:13,266 --> 00:07:14,534
REALLY CATASTROPHIC

201
00:07:14,534 --> 00:07:15,135
IF NOT

202
00:07:15,135 --> 00:07:16,369
FOR THE FACT

203
00:07:16,369 --> 00:07:19,572
THAT THE STALLED RNA POLYMERASE ITSELF,

204
00:07:19,839 --> 00:07:21,074
LIKE XPC,

205
00:07:21,074 --> 00:07:22,142
CAN ACTUALLY FUNCTION

206
00:07:22,142 --> 00:07:26,780
AS A SENSOR FOR THE REPAIR PATHWAY.

207
00:07:26,780 --> 00:07:28,782
AND THIS WAS THE INSIGHT OF PHIL HANAWALT

208
00:07:28,782 --> 00:07:31,151
IN THE MID 1980S.

209
00:07:31,151 --> 00:07:33,320
AND SO, WE NOW KNOW

210
00:07:33,320 --> 00:07:34,788
FROM THE WORK

211
00:07:34,788 --> 00:07:36,589
OF MANY PEOPLE IN THE FIELD,

212
00:07:36,589 --> 00:07:38,992
THE SORT OF OUTLINES OF THIS TC-NER

213
00:07:38,992 --> 00:07:39,993
THIS TRANSCRIPTION

214
00:07:39,993 --> 00:07:41,561
COUPLED NUCLEOTIDE EXCISION

215
00:07:41,561 --> 00:07:43,029
REPAIR PATHWAY.

216
00:07:43,029 --> 00:07:46,032
AND REALLY THE FIRST EVENT HERE

217
00:07:46,099 --> 00:07:47,233
IS THE RECRUITMENT

218
00:07:47,233 --> 00:07:50,136
OF AN ATPASE CALLED CSB.

219
00:07:50,136 --> 00:07:53,673
AND THIS PROTEIN IS MUTATED IN

220
00:07:53,673 --> 00:07:54,908
ANOTHER HUMAN DISORDER

221
00:07:54,908 --> 00:07:56,343
CALLED COCKAYNE SYNDROME,

222
00:07:56,343 --> 00:07:58,211
WHICH IS CHARACTERIZED BY,

223
00:07:58,211 --> 00:08:01,214
NEURODEGENERATION AND PROGERIA.

224
00:08:02,148 --> 00:08:03,683
AND WHAT CSB DOES

225
00:08:03,683 --> 00:08:05,652
IS IT ACTUALLY PUSHES ON THE BACK

226
00:08:05,652 --> 00:08:07,721
END OF RNA POLYMERASE

227
00:08:07,721 --> 00:08:09,856
AND TRIES TO MOVE IT PAST OBSTACLES.

228
00:08:09,856 --> 00:08:11,791
AND IF THAT FAILS, AS IN THE CASE OF

229
00:08:11,791 --> 00:08:15,462
VERY BULKY DNA LESIONS, CSB

230
00:08:15,462 --> 00:08:18,732
THEN NUCLEATES THE FORMATION OF A LARGE

231
00:08:18,865 --> 00:08:21,301
SO-CALLED TC-NER COMPLEX.

232
00:08:21,301 --> 00:08:24,070
AND SO THE NEXT PROTEIN TO LOAD IN

233
00:08:24,070 --> 00:08:25,538
THIS SEQUENCE IS AN E3

234
00:08:25,538 --> 00:08:28,541
UBIQUITIN LIGASE CALLED CRL4CSA,

235
00:08:28,875 --> 00:08:29,943
WHOSE KEY

236
00:08:29,943 --> 00:08:33,146
RECOGNITION SUBUNIT IS CALLED CSA.

237
00:08:33,279 --> 00:08:35,348
AND THAT'S ANOTHER ONE OF THE PROTEINS

238
00:08:35,348 --> 00:08:37,150
MUTATED IN THE COCKAYNE SYNDROME.

239
00:08:38,318 --> 00:08:39,753
AND CRL4

240
00:08:39,753 --> 00:08:41,287
DOES A FEW THINGS,

241
00:08:41,287 --> 00:08:43,289
ONE OF WHICH IS THAT IT ACTUALLY

242
00:08:43,289 --> 00:08:44,124
UBIQUITYLATES

243
00:08:44,124 --> 00:08:47,127
RNA POLYMERASE ON A SINGLE RESIDUE.

244
00:08:47,494 --> 00:08:48,828
AND THAT UBIQUITINATION

245
00:08:48,828 --> 00:08:51,097
EVENT IS CRUCIAL FOR REPAIR.

246
00:08:51,097 --> 00:08:52,632
ALTHOUGH WE DON'T REALLY UNDERSTAND

247
00:08:52,632 --> 00:08:55,301
THE MECHANISTIC BASIS FOR THAT.

248
00:08:55,301 --> 00:08:59,038
ANOTHER THING THAT THIS E3 LIGASE DOES

249
00:08:59,038 --> 00:09:00,673
IS TOGETHER WITH CSB, IT

250
00:09:00,673 --> 00:09:03,309
RECRUITS ADDITIONAL FACTORS,

251
00:09:03,309 --> 00:09:05,979
INCLUDING THE GIVES UVSSA PROTEIN

252
00:09:05,979 --> 00:09:07,847
TO COMPLETE THE FORMATION

253
00:09:07,847 --> 00:09:10,850
OF THIS SO-CALLED TC-NER COMPLEX.

254
00:09:11,618 --> 00:09:13,686
AND I'M GOING TO DENOTE IT HERE.

255
00:09:13,686 --> 00:09:14,387
I'M NOT GOING TO,

256
00:09:14,387 --> 00:09:18,158
CONTINUALLY DRAW THE ENTIRE E3 LIGASE.

257
00:09:18,958 --> 00:09:19,859
AND WE ACTUALLY HAVE

258
00:09:19,859 --> 00:09:21,027
A PRETTY GOOD UNDERSTANDING

259
00:09:21,027 --> 00:09:22,362
STRUCTURALLY OF WHAT

260
00:09:22,362 --> 00:09:25,498
THIS TC-NER COMPLEX LOOKS LIKE,

261
00:09:25,498 --> 00:09:27,100
LOOKS LIKE BASED ON BEAUTIFUL WORK

262
00:09:27,100 --> 00:09:29,502
FROM PATRICK KRAMER'S LABORATORY.

263
00:09:29,502 --> 00:09:32,672
AND REALLY, THE FUNCTION OF THIS COMPLEX,

264
00:09:32,672 --> 00:09:34,841
ULTIMATELY, IS TO RECRUIT

265
00:09:34,841 --> 00:09:37,410
THE TFIIH PROTEIN.

266
00:09:37,410 --> 00:09:40,180
AND AS WE SAW IN GLOBAL GENOMIC REPAIR

267
00:09:40,180 --> 00:09:42,048
AND TRANSCRIPTION COUPLED REPAIR,

268
00:09:42,048 --> 00:09:45,051
WHAT IT DOES IS IT LOCALLY OPENS THE DNA.

269
00:09:45,051 --> 00:09:47,954
THAT'S THE FUNCTION AGAIN OF XPB.

270
00:09:47,954 --> 00:09:49,422
AND THAT ALLOWS XPD

271
00:09:49,422 --> 00:09:52,525
TO ENGAGE WITH THE TRANSCRIBED STRAND

272
00:09:52,926 --> 00:09:55,628
AND XPD THEN TRANSLOCATE TOWARDS

273
00:09:55,628 --> 00:09:57,163
THE STALLED POLYMERASE

274
00:09:57,163 --> 00:09:58,665
LOOKING FOR DAMAGE.

275
00:09:58,665 --> 00:09:59,399
AND AGAIN,

276
00:09:59,399 --> 00:10:01,334
IF DAMAGE IS SENSED,

277
00:10:01,334 --> 00:10:02,602
THE WHOLE TFIIH

278
00:10:02,602 --> 00:10:04,337
COMPLEX STALLS AND RECRUITS

279
00:10:04,337 --> 00:10:05,205
THE NUCLEASES

280
00:10:05,205 --> 00:10:05,972
THAT THEN

281
00:10:05,972 --> 00:10:09,075
EXCISED THE LESION AND ALLOW GAP FILLING.

282
00:10:09,742 --> 00:10:13,079
A SECOND POSSIBLE FUNCTION OF THIS XPD

283
00:10:13,213 --> 00:10:16,850
DEPENDENT TRANSLOCATION IS THAT,

284
00:10:16,850 --> 00:10:18,251
IT PROBABLY.

285
00:10:18,251 --> 00:10:19,185
ALTHOUGH THIS HAS NEVER BEEN

286
00:10:19,185 --> 00:10:20,453
DIRECTLY OBSERVED, IT

287
00:10:20,453 --> 00:10:22,722
PROBABLY BACKTRACKS RNA POLYMERASE

288
00:10:22,722 --> 00:10:24,591
TO EXPOSE THE LESION

289
00:10:24,591 --> 00:10:26,459
AND EFFECTIVELY ALLOW

290
00:10:26,459 --> 00:10:28,127
THE DOWNSTREAM EVENTS.

291
00:10:29,162 --> 00:10:30,797
SO YOU CAN REALLY THINK OF,

292
00:10:30,797 --> 00:10:32,565
BOTH OF THESE PATHWAYS,

293
00:10:32,565 --> 00:10:33,433
BOTH BRANCHES,

294
00:10:33,433 --> 00:10:35,535
AS OCCURRING IN THREE BASIC STEPS.

295
00:10:35,535 --> 00:10:36,870
THE SENSING MECHANISM,

296
00:10:36,870 --> 00:10:39,205
WHICH IN THE CASE OF,

297
00:10:39,205 --> 00:10:40,406
GLOBAL GENOMIC REPAIR

298
00:10:40,406 --> 00:10:42,242
REALLY RELIES ON XPC

299
00:10:42,242 --> 00:10:43,243
AND IN THE CASE

300
00:10:43,243 --> 00:10:44,944
OF TRANSCRIPTION COUPLED REPAIR,

301
00:10:44,944 --> 00:10:46,980
REALLY RELIES ON RNA POLYMERASE.

302
00:10:46,980 --> 00:10:48,882
THEN THE VERIFICATION STEP,

303
00:10:48,882 --> 00:10:49,782
WHICH IS COMMON

304
00:10:49,782 --> 00:10:51,851
BETWEEN THE TWO AND DEPENDS ON TFIIH.

305
00:10:51,851 --> 00:10:52,752
AND THEN FINALLY ALSO

306
00:10:52,752 --> 00:10:53,520
THE COMMON

307
00:10:53,520 --> 00:10:56,523
EXCISION AND GAP FILLING EVENTS.

308
00:10:57,190 --> 00:11:00,527
SO, I HOPE THIS GIVES YOU,

309
00:11:00,527 --> 00:11:02,629
A PICTURE

310
00:11:02,629 --> 00:11:04,764
ABOUT WHERE WE ARE IN THE FIELD,

311
00:11:04,764 --> 00:11:06,266
BUT I WOULD LIKE TO POINT OUT

312
00:11:06,266 --> 00:11:07,700
THAT THERE ARE

313
00:11:07,700 --> 00:11:08,968
LOTS OF UNANSWERED

314
00:11:08,968 --> 00:11:09,769
QUESTIONS,

315
00:11:09,769 --> 00:11:12,772
PARTICULARLY ON THE TRANSCRIPTION COUPLED

316
00:11:13,006 --> 00:11:13,873
REPAIR SIDE.

317
00:11:13,873 --> 00:11:15,275
AND THAT'S BECAUSE THIS PATHWAY,

318
00:11:15,275 --> 00:11:17,076
UNLIKE THE GLOBAL GENOMIC PATHWAY,

319
00:11:17,076 --> 00:11:19,479
HAS NOT BEEN FULLY RECONSTITUTED,

320
00:11:19,479 --> 00:11:21,147
WITH PURIFIED COMPONENTS.

321
00:11:21,147 --> 00:11:23,883
AND IN FACT, THERE IS NO CELL FREE SYSTEM

322
00:11:23,883 --> 00:11:26,319
CURRENTLY, AT LEAST UNTIL,

323
00:11:26,319 --> 00:11:29,522
RECENTLY THAT SUPPORTS THIS PATHWAY.

324
00:11:29,522 --> 00:11:30,857
AND SO AS A RESULT,

325
00:11:30,857 --> 00:11:32,191
THERE ARE REALLY SOME

326
00:11:32,191 --> 00:11:33,459
FUNDAMENTAL QUESTIONS

327
00:11:33,459 --> 00:11:34,794
THAT REMAIN UNANSWERED.

328
00:11:34,794 --> 00:11:37,797
ONE OF THOSE IS HOW IS TFIIH

329
00:11:38,197 --> 00:11:40,366
ACTUALLY PROPERLY POSITIONED?

330
00:11:40,366 --> 00:11:41,434
REMEMBER THAT

331
00:11:41,434 --> 00:11:43,403
IN ORDER FOR PRODUCTIVE REPAIR

332
00:11:43,403 --> 00:11:45,004
TO OCCUR, TFIIH

333
00:11:45,004 --> 00:11:48,575
HAS TO BIND DOWNSTREAM OF THE STALLED

334
00:11:48,575 --> 00:11:49,442
RNA POLYMERASE,

335
00:11:49,442 --> 00:11:52,478
AND IT HAS TO SPECIFICALLY ENGAGE

336
00:11:52,478 --> 00:11:55,381
THROUGH XPD WITH THE TRANSCRIBED STRAND.

337
00:11:55,381 --> 00:11:55,815
THAT'S REALLY

338
00:11:55,815 --> 00:11:57,116
THE ONLY CONFIGURATION

339
00:11:57,116 --> 00:11:59,018
THAT WILL LEAD TO A

340
00:11:59,018 --> 00:12:00,453
PRODUCTIVE REPAIR OUTCOME.

341
00:12:00,453 --> 00:12:01,254
AND HOW

342
00:12:01,254 --> 00:12:02,922
TFIIH IS POSITIONED IN THAT

343
00:12:02,922 --> 00:12:04,824
MANNER IS NOT UNDERSTOOD.

344
00:12:04,824 --> 00:12:06,859
FURTHERMORE, WE REALLY UNDERSTAND

345
00:12:06,859 --> 00:12:08,561
LITTLE ABOUT THE

346
00:12:08,561 --> 00:12:13,232
DYNAMICES OF, OF RNA POLYMERASE.

347
00:12:13,232 --> 00:12:16,502
SO, IN THE COURSE OF REPAIR, IS IT,

348
00:12:16,736 --> 00:12:17,971
IS IT BACKTRACKED

349
00:12:17,971 --> 00:12:20,473
AND DOES IT RESTART OR IS IT REMOVED?

350
00:12:20,473 --> 00:12:22,709
THESE ARE SOME UNANSWERED QUESTIONS.

351
00:12:22,709 --> 00:12:24,043
AND ALSO

352
00:12:24,043 --> 00:12:26,012
WHAT IS THE ROLE OF RNA

353
00:12:26,012 --> 00:12:27,780
POLYMERASE UBIQUITINATION.

354
00:12:27,780 --> 00:12:28,982
AND FINALLY,

355
00:12:28,982 --> 00:12:30,683
GIVEN THAT THE PROCESS

356
00:12:30,683 --> 00:12:32,251
HAS NOT BEEN RECONSTITUTED

357
00:12:32,251 --> 00:12:34,053
TO PURIFIED COMPONENTS,

358
00:12:34,053 --> 00:12:37,123
ARE WE ACTUALLY MISSING ANY KEY FACTORS.

359
00:12:37,123 --> 00:12:38,725
AND MIGHT SUCH FACTORS SHED

360
00:12:38,725 --> 00:12:41,728
LIGHT ON ON THE ABOVE QUESTIONS?

361
00:12:41,728 --> 00:12:43,062
SO TODAY I'M GOING TO TELL YOU

362
00:12:43,062 --> 00:12:44,397
ABOUT THE WORK OF A REALLY,

363
00:12:44,397 --> 00:12:45,732
REALLY TALENTED,

364
00:12:45,732 --> 00:12:46,666
POSTDOCTORAL FELLOW,

365
00:12:46,666 --> 00:12:47,500
TYCHO MEVISSEN,

366
00:12:47,500 --> 00:12:49,736
WHO ACTUALLY IS ON THE JOB MARKET.

367
00:12:49,736 --> 00:12:52,739
AND WHAT TYCHO DID WAS TO SAY,

368
00:12:53,172 --> 00:12:56,376
LET'S TRY TO ACTUALLY DEVELOP

369
00:12:56,376 --> 00:12:57,577
A CELL-FREE SYSTEM

370
00:12:57,577 --> 00:13:00,480
THAT SUPPORTS EUKARYOTIC TC-NER.

371
00:13:00,480 --> 00:13:02,482
THERE HAVE BEEN SEVERAL ATTEMPTS

372
00:13:02,482 --> 00:13:05,485
OVER THE YEARS TO ACHIEVE THIS,

373
00:13:05,551 --> 00:13:07,220
AND, UNFORTUNATELY,

374
00:13:07,220 --> 00:13:07,887
NONE OF THEM

375
00:13:07,887 --> 00:13:09,589
HAVE BEEN SUCCESSFUL IN AND

376
00:13:09,589 --> 00:13:10,923
AND THAT'S ACTUALLY PRETTY CLEAR

377
00:13:10,923 --> 00:13:11,824
FROM THE FACT

378
00:13:11,824 --> 00:13:13,660
THAT NONE OF THE CELL-FREE SYSTEMS

379
00:13:13,660 --> 00:13:14,761
THAT HAVE BEEN,

380
00:13:14,761 --> 00:13:16,562
PROPOSED OR DOCUMENTED

381
00:13:16,562 --> 00:13:19,599
ACTUALLY REALLY DEPEND ON THE CORE

382
00:13:19,599 --> 00:13:20,266
TC-NER

383
00:13:20,266 --> 00:13:24,003
FACTORS SUCH AS CSB AND CRL4CSA.

384
00:13:24,904 --> 00:13:27,907
SO TYCHO REASONED THAT

385
00:13:28,041 --> 00:13:30,043
PERHAPS GIVEN

386
00:13:30,043 --> 00:13:32,111
THE EXTRAORDINARY POWER

387
00:13:32,111 --> 00:13:34,313
OF XENOPUS EGG EXTRACTS, TO RECAPITULATE

388
00:13:35,281 --> 00:13:38,084
SO MANY GENOME MAINTENANCE PATHWAYS

389
00:13:38,084 --> 00:13:40,019
INCLUDING CHROMATIN ASSEMBLY,

390
00:13:40,019 --> 00:13:41,754
CHECKPOINT SIGNALING,

391
00:13:41,754 --> 00:13:43,456
REPLICATION AND ALL OF THESE REPAIR

392
00:13:43,456 --> 00:13:44,590
PATHWAYS, MIGHT

393
00:13:44,590 --> 00:13:47,226
IT ALSO BE ABLE TO SUPPORT TC-NER?

394
00:13:47,226 --> 00:13:49,028
THAT WAS REALLY,

395
00:13:49,028 --> 00:13:52,331
THE SORT OF THE OVERARCHING RATIONALE

396
00:13:52,331 --> 00:13:53,166
OF HIS PROJECT.

397
00:13:53,166 --> 00:13:55,635
THERE WAS ONLY ONE SMALL DETAIL,

398
00:13:55,635 --> 00:13:57,270
AND THAT IS THAT FROG

399
00:13:57,270 --> 00:13:59,005
EGG EXTRACTS, AT LEAST CONVENTIONAL

400
00:13:59,005 --> 00:14:02,775
EGG EXTRACTS, WERE KNOWN TO BE INACTIVE

401
00:14:02,909 --> 00:14:04,110
FOR TRANSCRIPTION.

402
00:14:04,110 --> 00:14:05,712
SO THE FIRST THING THAT TYCHO

403
00:14:05,712 --> 00:14:08,715
HAD TO DO TO TRY TO ACTIVATE TC-NER IN

404
00:14:08,715 --> 00:14:10,483
THIS SYSTEM WAS TO ACTIVATE

405
00:14:10,483 --> 00:14:11,818
TRANSCRIPTION ITSELF.

406
00:14:11,818 --> 00:14:13,252
AND THE WAY HE WENT ABOUT DOING

407
00:14:13,252 --> 00:14:16,489
THIS WAS TO GENERATE A PLASMID

408
00:14:16,489 --> 00:14:19,492
THAT CONTAINS, A VERY STRONG.

409
00:14:21,861 --> 00:14:22,695
SOMEHOW MY,

410
00:14:22,695 --> 00:14:24,731
MY MOUSE IS NOT RESPONDING PROPERLY.

411
00:14:24,731 --> 00:14:27,200
HERE. HOLD ON ONE SECOND. LET'S SEE.

412
00:14:27,200 --> 00:14:29,268
MAYBE THERE'S SOME JUNK ON MY.

413
00:14:29,268 --> 00:14:31,037
I'M, INTERESTING.

414
00:14:31,037 --> 00:14:32,739
HOLD ON ONE SECOND.

415
00:14:38,344 --> 00:14:39,679
OKAY.

416
00:14:39,679 --> 00:14:40,880
THIS IS FASCINATING.

417
00:14:40,880 --> 00:14:43,950
MY MOUSE DECIDED TO STOP WORKING.

418
00:14:45,918 --> 00:14:46,285
OKAY.

419
00:14:46,285 --> 00:14:49,288
WELL, SO,

420
00:14:49,455 --> 00:14:52,291
HE USED A PLASMID,

421
00:14:52,291 --> 00:14:54,861
CONTAINING A VERY STRONG CORE PROMOTER

422
00:14:54,861 --> 00:14:56,929
AS WELL AS A FEW UPSTREAM

423
00:14:56,929 --> 00:14:58,631
ACTIVATING SEQUENCES.

424
00:14:58,631 --> 00:15:00,333
AND HE ADDED THIS INTO EGG

425
00:15:00,333 --> 00:15:01,467
EXTRACT TOGETHER

426
00:15:01,467 --> 00:15:03,402
WITH A POTENT TRANSCRIPTIONAL ACTIVATOR

427
00:15:03,402 --> 00:15:05,972
GAL4VP64.

428
00:15:05,972 --> 00:15:09,475
AND INSTEAD OF USING CONVENTIONAL TOTAL

429
00:15:09,475 --> 00:15:10,777
EGG LYSATES,

430
00:15:10,777 --> 00:15:14,514
WHICH OTHERS HAD TRIED BEFORE AND REALLY

431
00:15:14,814 --> 00:15:16,749
DIDN'T LEAD TO ANY SIGNIFICANT

432
00:15:16,749 --> 00:15:18,284
TRANSCRIPTIONAL ACTIVATION,

433
00:15:18,284 --> 00:15:19,452
WHAT HE DID

434
00:15:19,452 --> 00:15:22,088
INSTEAD WAS TO USE A CONCENTRATED

435
00:15:22,088 --> 00:15:23,222
NUCLEARPLASMIC

436
00:15:23,222 --> 00:15:25,291
EXTRACT THAT IS HIGHLY ENRICHED

437
00:15:25,291 --> 00:15:27,393
IN NUCLEAR PROTEINS,

438
00:15:27,393 --> 00:15:29,395
INCLUDING RNA POLYMERASE, MEDIATOR,

439
00:15:29,395 --> 00:15:30,329
AND ALL OF THESE THINGS

440
00:15:30,329 --> 00:15:31,197
THAT THAT ARE NEEDED

441
00:15:31,197 --> 00:15:32,532
TO ACTIVATE TRANSCRIPTION.

442
00:15:32,532 --> 00:15:34,300
AND WHEN HE DID THIS

443
00:15:34,300 --> 00:15:37,170
AND ALSO INCLUDED RADIOACTIVE UTP,

444
00:15:37,170 --> 00:15:39,939
HE SAW A VERY POTENT TRANSCRIPTION.

445
00:15:39,939 --> 00:15:41,440
SO THIS WAS GREAT.

446
00:15:41,440 --> 00:15:44,811
AND, AND MEANT THAT HE HAD OVERCOME

447
00:15:44,811 --> 00:15:46,746
THE FIRST HURDLE TOWARDS

448
00:15:46,746 --> 00:15:48,681
RECAPITULATING TC-NER.

449
00:15:48,681 --> 00:15:49,849
AND I SHOULD NOTE THAT ACTUALLY,

450
00:15:49,849 --> 00:15:51,784
MY FORMER POSTDOC, DAVID LONG,

451
00:15:51,784 --> 00:15:54,020
WHO HAS HIS OWN LABORATORY

452
00:15:54,020 --> 00:15:55,321
IN SOUTH CAROLINA,

453
00:15:55,321 --> 00:15:56,923
ALSO HAS ACTIVATED

454
00:15:56,923 --> 00:15:58,691
TRANSCRIPTION IN FROG EGG EXTRACTS.

455
00:15:58,691 --> 00:15:59,292
ALTHOUGH,

456
00:15:59,292 --> 00:16:00,593
HIS APPROACH INVOLVES THE

457
00:16:00,593 --> 00:16:02,328
USE OF ENDOGENOUS PROMOTERS.

458
00:16:03,329 --> 00:16:05,164
SO, THE GOAL HERE,

459
00:16:05,164 --> 00:16:08,901
OF COURSE, WAS TO RECAPITULATE TC-NER.

460
00:16:08,901 --> 00:16:10,503
SO THE NEXT THING THAT TYCHO DID

461
00:16:10,503 --> 00:16:13,573
WAS TO ENGINEER A CLASSIC NUCLEOTIDE

462
00:16:13,573 --> 00:16:17,410
EXCISION REPAIR SUBSTRATE, A 1,3-GTG

463
00:16:17,510 --> 00:16:18,911
INTRASTRAND

464
00:16:18,911 --> 00:16:21,914
CROSSLINK INTO THE TEMPLATE STRAND.

465
00:16:22,014 --> 00:16:23,749
AND AS YOU CAN SEE FROM THE GEL

466
00:16:23,749 --> 00:16:25,218
OVER ON THE RIGHT,

467
00:16:25,218 --> 00:16:27,253
THIS LED TO VERY POTENT

468
00:16:27,253 --> 00:16:29,322
TRANSCRIPTIONAL ARREST.

469
00:16:29,322 --> 00:16:30,990
AND SO NOW THE QUESTION BECAME

470
00:16:30,990 --> 00:16:33,059
WHETHER ANY OF THESE LESIONS

471
00:16:33,059 --> 00:16:34,260
ARE REPAIRED

472
00:16:34,260 --> 00:16:37,129
IN A TRANSCRIPTION COUPLED MANNER.

473
00:16:37,129 --> 00:16:39,031
EXCUSE ME.

474
00:16:39,031 --> 00:16:41,234
AND TO ADDRESS THIS,

475
00:16:41,234 --> 00:16:43,102
TYCHO USED A SIMPLE TRICK

476
00:16:43,102 --> 00:16:45,938
WHICH WAS TO ASK

477
00:16:45,938 --> 00:16:48,641
WHETHER A PML1 RESTRICTION SITE

478
00:16:48,641 --> 00:16:49,775
THAT COINCIDES

479
00:16:49,775 --> 00:16:52,979
WITH THE LESION IS REGENERATED.

480
00:16:53,145 --> 00:16:55,781
SO THE INPUT PLASMID CANNOT BE CUT

481
00:16:55,781 --> 00:16:57,283
WITH THIS ENZYME DUE

482
00:16:57,283 --> 00:16:58,651
TO THE PRESENCE OF THE LESION.

483
00:16:58,651 --> 00:17:00,586
AND IF ANY ERROR-FREE REPAIR

484
00:17:00,586 --> 00:17:01,787
SHOULD OCCUR,

485
00:17:01,787 --> 00:17:04,423
THE PLASMID WILL BECOME CLEAVABLE.

486
00:17:05,424 --> 00:17:08,427
SO WHEN HE FIRST

487
00:17:08,728 --> 00:17:10,696
USED THIS ASSAY TO ASK

488
00:17:10,696 --> 00:17:12,765
WHETHER ANY REPAIR IS GOING ON,

489
00:17:12,765 --> 00:17:14,700
HE ACTUALLY COULD SEE

490
00:17:14,700 --> 00:17:17,970
THAT ABOUT 10 TO 30%,

491
00:17:17,970 --> 00:17:19,005
DEPENDING ON THE DAY

492
00:17:19,005 --> 00:17:21,941
AND THE EXTRACT OF THE INPUT PLASMID

493
00:17:21,941 --> 00:17:24,043
WAS UNDERGOING REPAIR.

494
00:17:24,043 --> 00:17:27,847
HOWEVER, HE THEN VERY QUICKLY UNDERSTOOD

495
00:17:28,214 --> 00:17:31,050
THAT THIS IS DUE TO GLOBAL GENOMIC REPAIR

496
00:17:31,050 --> 00:17:33,252
BECAUSE IF HE DEPLETED XPC,

497
00:17:33,252 --> 00:17:34,620
THE INITIATOR OF THE GLOBAL

498
00:17:34,620 --> 00:17:35,922
GENOMIC PATHWAY

499
00:17:35,922 --> 00:17:37,723
REPAIR WAS ELIMINATED

500
00:17:37,723 --> 00:17:38,991
EVEN UNDER CONDITIONS

501
00:17:38,991 --> 00:17:41,994
WHERE HE WAS ACTIVATING TRANSCRIPTION.

502
00:17:42,428 --> 00:17:44,797
SO IT TOOK US A LITTLE WHILE,

503
00:17:44,797 --> 00:17:45,631
BUT EVENTUALLY

504
00:17:45,631 --> 00:17:48,901
WE REALIZED THAT THIS IS QUITE POSSIBLY

505
00:17:48,901 --> 00:17:50,436
OR PROBABLY DUE,

506
00:17:50,436 --> 00:17:53,439
AT LEAST IN PART TO THE FACT THAT

507
00:17:53,606 --> 00:17:57,877
THE CONCENTRATION OF CLASSIC TC-NER

508
00:17:57,877 --> 00:17:58,911
PROTEINS,

509
00:17:58,911 --> 00:18:00,446
SUCH AS CSB AND CRL4CSA

510
00:18:00,446 --> 00:18:03,449
IN EGG EXTRACT IS VERY LOW.

511
00:18:04,517 --> 00:18:07,153
SO THEREFORE, TYCHO GENERATED

512
00:18:07,153 --> 00:18:11,791
A COCKTAIL OF KNOWN AND SUSPECTED

513
00:18:11,791 --> 00:18:15,061
TC-NER FACTORS THAT'S SHOWN, DOWN HERE.

514
00:18:15,561 --> 00:18:19,298
AND VERY EXCITINGLY, WHEN HE SUPPLEMENTED

515
00:18:19,398 --> 00:18:21,400
THE EGG EXTRACTS WITH THIS COCKTAIL,

516
00:18:21,400 --> 00:18:23,135
THIS SO-CALLED TC-NER COCKTAIL,

517
00:18:23,135 --> 00:18:26,072
HE NOW SAW A MASSIVE BOOST

518
00:18:26,072 --> 00:18:28,407
IN THE EFFICIENCY OF REPAIR,

519
00:18:28,407 --> 00:18:29,575
EVEN UNDER CONDITIONS

520
00:18:29,575 --> 00:18:32,011
WHERE GLOBAL GENOMIC REPAIR

521
00:18:32,011 --> 00:18:33,612
WAS STILL INHIBITED

522
00:18:33,612 --> 00:18:36,248
DUE TO THE REMOVAL OF XPC.

523
00:18:36,248 --> 00:18:38,117
AND HE GOT REALLY EXCITED

524
00:18:38,117 --> 00:18:40,853
WHEN HE FOUND THAT IF HE ALSO NOW

525
00:18:40,853 --> 00:18:42,588
INHIBITED TRANSCRIPTION

526
00:18:42,588 --> 00:18:43,990
USING ALPHA-AMANITIN,

527
00:18:43,990 --> 00:18:46,158
THAT REPAIR CAME BACK DOWN AGAIN.

528
00:18:46,158 --> 00:18:47,994
SO THIS REALLY SUGGESTED

529
00:18:47,994 --> 00:18:50,997
THAT FOR THE FIRST TIME HE COULD OBSERVE,

530
00:18:51,197 --> 00:18:54,367
BONA FIDE, TRANSCRIPTION COUPLED REPAIR.

531
00:18:54,367 --> 00:18:56,102
AND HE WENT ON AND EXPERIMENTS

532
00:18:56,102 --> 00:18:57,403
THAT I WON'T SHOW YOU

533
00:18:57,403 --> 00:18:58,704
TO DEMONSTRATE

534
00:18:58,704 --> 00:19:01,640
THAT THIS TRANSCRIPTION DEPENDENT

535
00:19:01,640 --> 00:19:03,342
REPAIR SIGNAL

536
00:19:03,342 --> 00:19:05,911
THAT HE SEES IS ABSOLUTELY DEPENDENT

537
00:19:05,911 --> 00:19:10,282
ON ALL FOUR OF THE CANONICAL TC-NER FACTORS.

538
00:19:10,282 --> 00:19:11,217
AND SO THAT REALLY

539
00:19:11,217 --> 00:19:13,152
CONVINCED US THAT FOR THE FIRST TIME,

540
00:19:13,152 --> 00:19:15,654
WE HAD BEEN ABLE TO RECAPITULATE

541
00:19:15,654 --> 00:19:19,058
BONA FIDE A TC-NER IN THE TEST TUBE.

542
00:19:19,058 --> 00:19:20,593
AND THIS NOW SET THE STAGE

543
00:19:20,593 --> 00:19:21,961
TO REALLY EMBARK

544
00:19:21,961 --> 00:19:25,331
ON A DETAILED MECHANISTIC ANALYSIS OF,

545
00:19:25,598 --> 00:19:26,599
OF THIS PATHWAY

546
00:19:26,599 --> 00:19:27,867
AND TO TRY TO ADDRESS SOME

547
00:19:27,867 --> 00:19:29,201
OF THE UNANSWERED QUESTIONS.

548
00:19:30,703 --> 00:19:32,538
SO THE FIRST,

549
00:19:32,538 --> 00:19:37,643
QUESTION THAT HE TACKLED WAS TO ADDRESS

550
00:19:37,643 --> 00:19:39,478
THE FUNCTION OF AN OBSCURE

551
00:19:39,478 --> 00:19:41,514
PROTEIN CALLED STK19

552
00:19:41,514 --> 00:19:44,917
THAT STANDS FOR SERINE/THREONINE KINASE 19

553
00:19:45,317 --> 00:19:48,320
AND IS A TOTAL MISNOMER BECAUSE, IN FACT,

554
00:19:48,687 --> 00:19:50,322
THE PROTEIN CONSISTS

555
00:19:50,322 --> 00:19:53,059
OF A SERIES OF WINGED HELIX DOMAINS.

556
00:19:53,059 --> 00:19:54,527
IT HAS NO KINASE FOLD.

557
00:19:54,527 --> 00:19:57,296
IT HAS NO DETECTABLE KINASE ACTIVITY.

558
00:19:57,296 --> 00:19:58,064
HOWEVER, WHAT'S

559
00:19:58,064 --> 00:19:59,365
INTERESTING ABOUT THE PROTEIN

560
00:19:59,365 --> 00:20:01,100
IS THAT IT HAD BEEN CONNECTED

561
00:20:01,100 --> 00:20:03,102
TO TC-NER,

562
00:20:03,102 --> 00:20:04,370
BY VARIOUS GROUPS

563
00:20:04,370 --> 00:20:06,639
IN DIFFERENT INDIRECT WAYS,

564
00:20:06,639 --> 00:20:07,907
THE MOST INTERESTING

565
00:20:07,907 --> 00:20:10,142
OF WHICH PROBABLY IS THAT

566
00:20:10,142 --> 00:20:11,677
IT IS REQUIRED

567
00:20:11,677 --> 00:20:12,611
FOR THE RESUMPTION

568
00:20:12,611 --> 00:20:16,449
OF TRANSCRIPTION AFTER, A UV

569
00:20:16,449 --> 00:20:19,452
INSULT OR CELLULAR UV EXPOSURE

570
00:20:20,352 --> 00:20:21,454
AS SEEN

571
00:20:21,454 --> 00:20:24,857
FOR ESSENTIALLY ALL OTHER CORE

572
00:20:24,857 --> 00:20:26,125
TC-NER FACTORS.

573
00:20:26,125 --> 00:20:26,926
AND IN FACT,

574
00:20:26,926 --> 00:20:28,894
WHEN TYCHO DEPLETED THE PROTEIN

575
00:20:28,894 --> 00:20:30,296
FROM HIS CELL FREE SYSTEM,

576
00:20:30,296 --> 00:20:33,466
HE SAW A COMPLETE ABOLITION OF CELL FREE,

577
00:20:33,833 --> 00:20:35,167
REPAIR,

578
00:20:35,167 --> 00:20:37,036
WHICH HE COULD RESCUE BY ADDING BACK

579
00:20:37,036 --> 00:20:39,705
RECOMBINANT STK19.

580
00:20:39,705 --> 00:20:40,706
SO THIS WAS GREAT.

581
00:20:40,706 --> 00:20:42,074
REALLY ARGUED THAT

582
00:20:42,074 --> 00:20:46,011
THAT, STK19 IS A TRUE

583
00:20:47,046 --> 00:20:48,814
TC-NER REPAIR FACTOR.

584
00:20:48,814 --> 00:20:50,416
AND IT RAISED THE QUESTION

585
00:20:50,416 --> 00:20:53,052
HOW STK19 FUNCTIONS.

586
00:20:53,052 --> 00:20:56,355
NOW, GIVEN THE ABSENCE OF ANY,

587
00:20:58,824 --> 00:20:59,492
CLEARLY

588
00:20:59,492 --> 00:21:02,495
IDENTIFIABLE ENZYMATIC ACTIVITY,

589
00:21:02,661 --> 00:21:06,232
THE NULL HYPOTHESIS BECAME THAT STK19

590
00:21:06,899 --> 00:21:08,067
FUNCTIONS IN TC-NER

591
00:21:08,067 --> 00:21:11,303
ARE BY INTERACTING WITH OTHER PROTEINS.

592
00:21:11,937 --> 00:21:16,809
AND SO, AS YOU ALL KNOW, IDENTIFYING

593
00:21:16,809 --> 00:21:18,344
FUNCTIONALLY RELEVANT PROTEIN

594
00:21:18,344 --> 00:21:21,580
PROTEIN INTERACTIONS IS VERY DIFFICULT.

595
00:21:21,747 --> 00:21:24,049
TYPICALLY ONE MIGHT DO AN

596
00:21:24,049 --> 00:21:25,651
IP MASS SPEC EXPERIMENT

597
00:21:25,651 --> 00:21:27,253
AND LOOK FOR

598
00:21:27,253 --> 00:21:28,754
THINGS THAT CO-IP

599
00:21:28,754 --> 00:21:30,689
WITH, WITH ONE'S PROTEIN OF INTEREST.

600
00:21:30,689 --> 00:21:32,358
BUT THEN THERE'S A LONG ROAD,

601
00:21:32,358 --> 00:21:34,360
IN TERMS OF

602
00:21:34,360 --> 00:21:37,663
VALIDATING WHICH OF THE MANY PROTEINS

603
00:21:37,663 --> 00:21:39,498
ONE ONE RECOVERS

604
00:21:39,498 --> 00:21:41,767
AS ACTUALLY BEING FUNCTIONALLY RELEVANT.

605
00:21:41,767 --> 00:21:42,935
AND THAT OFTEN INVOLVES

606
00:21:42,935 --> 00:21:44,370
AN ELABORATE STRUCTURE

607
00:21:44,370 --> 00:21:45,905
FUNCTION ANALYSIS,

608
00:21:45,905 --> 00:21:48,374
DOMAIN MAPPING AND, AND SO FORTH.

609
00:21:48,374 --> 00:21:50,342
AND THIS CAN REALLY BE A VERY,

610
00:21:50,342 --> 00:21:53,345
VERY LONG AND DIFFICULT PROCESS.

611
00:21:53,612 --> 00:21:56,582
AND SO OVER THE LAST COUPLE OF YEARS,

612
00:21:56,582 --> 00:21:58,918
STIMULATED BY THE STRUCTURE

613
00:21:58,918 --> 00:22:00,553
PREDICTION REVOLUTION,

614
00:22:00,553 --> 00:22:02,688
WE ASKED WHETHER WE CAN

615
00:22:02,688 --> 00:22:03,556
MAKE SORT OF A HAIL

616
00:22:03,556 --> 00:22:04,490
MARY PLAY

617
00:22:04,490 --> 00:22:06,825
AROUND THE CONVENTIONAL APPROACHES

618
00:22:06,825 --> 00:22:08,294
AND ACTUALLY HARNESS

619
00:22:08,294 --> 00:22:09,028
THE STRUCTURE

620
00:22:09,028 --> 00:22:12,198
PREDICTION REVOLUTION TO IDENTIFY

621
00:22:12,198 --> 00:22:13,933
FUNCTIONALLY RELEVANT

622
00:22:13,933 --> 00:22:16,569
PROTEIN-PROTEIN INTERACTIONS AND

623
00:22:16,569 --> 00:22:17,970
THIS IS THE WORK OF A VERY TALENTED

624
00:22:17,970 --> 00:22:20,439
GRADUATE STUDENT, ERNST SCHMID.

625
00:22:20,439 --> 00:22:21,473
AND ESSENTIALLY

626
00:22:21,473 --> 00:22:24,476
WHAT HE DID WAS TO USE THE,

627
00:22:25,144 --> 00:22:26,412
DEEP LEARNING STRUCTURE

628
00:22:26,412 --> 00:22:27,580
PREDICTION ALGORITHM.

629
00:22:27,580 --> 00:22:30,583
ALPHAFOLD-MULTIMER, A COUSIN OF ALPHAFOLD

630
00:22:30,883 --> 00:22:35,621
WHO'S REALLY, DEDICATED FUNCTION

631
00:22:35,621 --> 00:22:37,890
IS TO SOLVE THE STRUCTURE

632
00:22:37,890 --> 00:22:40,292
OF KNOWN PROTEIN-PROTEIN INTERACTIONS.

633
00:22:40,292 --> 00:22:41,860
AND THAT'S OBVIOUSLY SUPER USEFUL.

634
00:22:41,860 --> 00:22:42,194
AND THAT'S

635
00:22:42,194 --> 00:22:44,863
WHAT ALPHAFOLD-MULTIMER WAS DESIGNED FOR.

636
00:22:44,863 --> 00:22:48,167
BUT WHAT ERNST REALIZED, TOGETHER WITH,

637
00:22:48,167 --> 00:22:50,169
SEVERAL OTHER GROUPS LIKE THE BAKER LAB

638
00:22:50,169 --> 00:22:51,770
THE ELOFFSON LAB AND ALSO DAN

639
00:22:51,770 --> 00:22:52,871
DUROCHERS LAB,

640
00:22:52,871 --> 00:22:54,440
WAS THAT

641
00:22:54,440 --> 00:22:56,075
PERHAPS ONE COULD USE

642
00:22:56,075 --> 00:22:58,577
THIS STRUCTURE PREDICTION

643
00:22:58,577 --> 00:23:00,179
APPROACH TO ACTUALLY IDENTIFY

644
00:23:00,179 --> 00:23:01,814
NEW PROTEIN-PROTEIN INTERACTIONS.

645
00:23:01,814 --> 00:23:03,082
AND THE WAY THAT ERNST DOES

646
00:23:03,082 --> 00:23:05,351
THIS IS TO TAKE A BAIT PROTEIN

647
00:23:05,351 --> 00:23:08,254
AND THEN FOLD IT WITH,

648
00:23:08,254 --> 00:23:09,455
A LARGE NUMBER

649
00:23:09,455 --> 00:23:11,557
OF SO-CALLED PREY PROTEINS

650
00:23:11,557 --> 00:23:12,858
AND BASICALLY GENERATE

651
00:23:12,858 --> 00:23:15,127
BINARY STRUCTURE PREDICTIONS

652
00:23:15,127 --> 00:23:16,729
OF EACH PAIR USING ALPHAFOLD-MULTIMER,

653
00:23:18,197 --> 00:23:21,200
AND THEN, USE THE

654
00:23:21,700 --> 00:23:24,470
CONFIDENCE METRICS THAT ALPHAFOLD

655
00:23:24,470 --> 00:23:27,473
PUTS OUT TO PRIORITIZE THESE

656
00:23:27,873 --> 00:23:30,242
ALL OF THESE PREDICTIONS.

657
00:23:30,242 --> 00:23:32,645
AND ESSENTIALLY TO ASK

658
00:23:32,645 --> 00:23:33,445
WHICH OF THESE

659
00:23:33,445 --> 00:23:34,613
IS MOST LIKELY

660
00:23:34,613 --> 00:23:36,015
TO BE A FUNCTIONALLY

661
00:23:36,015 --> 00:23:37,616
RELEVANT INTERACTION?

662
00:23:37,616 --> 00:23:38,484
AND I'M NOT GOING TO GO

663
00:23:38,484 --> 00:23:40,719
INTO THE TECHNICAL DETAILS OF

664
00:23:40,719 --> 00:23:44,189
THE ALPHAFOLD CONFIDENCE METRICS

665
00:23:44,189 --> 00:23:45,124
THAT ARE LISTED HERE,

666
00:23:45,124 --> 00:23:48,127
THINGS LIKE IPTM, PLDDT AND SO FORTH

667
00:23:48,460 --> 00:23:50,562
JUST BECAUSE IT DOESN'T

668
00:23:50,562 --> 00:23:52,431
ADD MUCH TO THE DISCUSSION.

669
00:23:52,431 --> 00:23:55,434
BUT I WILL TELL YOU THAT FOR THESE

670
00:23:55,901 --> 00:23:56,268
AT LEAST

671
00:23:56,268 --> 00:23:57,703
FOR THESE LIMITED SCREENS,

672
00:23:57,703 --> 00:24:00,039
THESE HAVE BEEN REALLY QUITE USEFUL

673
00:24:00,039 --> 00:24:00,973
IN IDENTIFYING

674
00:24:00,973 --> 00:24:03,909
FUNCTIONALLY RELEVANT INTERACTIONS.

675
00:24:03,909 --> 00:24:07,846
AND SO SPURRED BY SOME EARLY SUCCESSES

676
00:24:07,846 --> 00:24:09,448
THAT I'LL TELL YOU ABOUT IN A MOMENT,

677
00:24:09,448 --> 00:24:11,583
ULTIMATELY, WE ACTUALLY WENT AHEAD

678
00:24:11,583 --> 00:24:12,818
AND FOLDED

679
00:24:12,818 --> 00:24:15,888
300 CORE GENOME MAINTENANCE PROTEINS

680
00:24:16,021 --> 00:24:16,789
WITH EACH OTHER,

681
00:24:16,789 --> 00:24:19,758
GENERATING A MATRIX OF ABOUT 40,000 PAIRS

682
00:24:19,758 --> 00:24:21,627
THAT WE MADE FREELY AVAILABLE

683
00:24:21,627 --> 00:24:22,528
AT THIS WEBSITE

684
00:24:22,528 --> 00:24:23,996
CALLED PREDICTOMES.ORG,

685
00:24:23,996 --> 00:24:26,465
WHICH HOPEFULLY MANY OF YOU HAVE SEEN

686
00:24:26,465 --> 00:24:30,636
AND AND AND ARE ACTUALLY USING AND

687
00:24:32,204 --> 00:24:32,838
I WILL

688
00:24:32,838 --> 00:24:33,906
TRY TO ILLUSTRATE THE

689
00:24:33,906 --> 00:24:36,108
POWER OF THIS RESOURCE

690
00:24:36,108 --> 00:24:37,309
USING ONE EXAMPLE

691
00:24:37,309 --> 00:24:39,978
BEFORE I COME BACK TO THE STK19 STORY.

692
00:24:39,978 --> 00:24:40,746
SO ONE OF THE

693
00:24:40,746 --> 00:24:42,981
THE EARLIEST SUCCESS STORIES

694
00:24:42,981 --> 00:24:46,385
FOR US WAS USING STRUCTURE PREDICTION

695
00:24:46,385 --> 00:24:47,119
TO IDENTIFY

696
00:24:47,119 --> 00:24:48,787
FUNCTIONALLY RELEVANT PARTNERS

697
00:24:48,787 --> 00:24:49,822
AND THEREFORE

698
00:24:49,822 --> 00:24:51,890
THE FUNCTION OF A PROTEIN CALLED DONSON.

699
00:24:51,890 --> 00:24:53,192
AND SO THAT PROTEIN

700
00:24:53,192 --> 00:24:54,426
HAD BEEN IMPLICATED

701
00:24:54,426 --> 00:24:55,928
IN SOME ASPECT OF DNA

702
00:24:55,928 --> 00:24:57,796
REPLICATION PREVIOUSLY.

703
00:24:57,796 --> 00:24:59,998
BUT WHAT IT REALLY DOES

704
00:24:59,998 --> 00:25:02,634
WAS QUITE OBSCURE.

705
00:25:02,634 --> 00:25:05,571
AND SO IN THIS STRUCTURE

706
00:25:05,571 --> 00:25:07,039
PREDICTION RESOURCE

707
00:25:07,039 --> 00:25:08,006
THAT WE'VE CREATED,

708
00:25:08,006 --> 00:25:11,210
YOU CAN, CLICK ON YOUR FAVORITE PROTEIN

709
00:25:11,210 --> 00:25:14,713
AND THEN RETRIEVE A PRIORITIZED LIST

710
00:25:14,713 --> 00:25:17,683
BASED ON THE ALPHAFOLD CONFIDENCE METRICS

711
00:25:18,283 --> 00:25:22,654
OF THE LIKELY OR POSSIBLE INTERACTORS.

712
00:25:22,654 --> 00:25:24,656
AND IN THE CASE OF DONSON,

713
00:25:24,656 --> 00:25:27,192
THE RESULTS WERE REALLY QUITE STRIKING.

714
00:25:27,192 --> 00:25:30,195
SO IN THIS LIST OF 300 PROTEINS,

715
00:25:30,195 --> 00:25:33,165
THE TOP 5 OR 6 PROTEINS,

716
00:25:33,165 --> 00:25:36,034
ALL INVOLVED FACTORS

717
00:25:36,034 --> 00:25:38,170
THAT HAD PREVIOUSLY BEEN IMPLICATED

718
00:25:38,170 --> 00:25:39,772
IN THE ASSEMBLY OF THE REPLICATIVE

719
00:25:39,772 --> 00:25:41,440
CMG HELICASE,

720
00:25:41,440 --> 00:25:43,041
WHICH CONSISTS OF THE MCM

721
00:25:43,041 --> 00:25:44,576
TWO THROUGH SEVEN ATPASE,

722
00:25:44,576 --> 00:25:45,911
AND THEN THESE TWO ATPASE

723
00:25:45,911 --> 00:25:49,181
COFACTORS GINS AND CDC45.

724
00:25:49,815 --> 00:25:52,050
SO WHAT ALPHAFOLD SUGGESTED

725
00:25:52,050 --> 00:25:54,052
WAS THAT DANCING INTERACTS WITH TOPB1

726
00:25:54,052 --> 00:25:55,854
FACTOR,

727
00:25:55,854 --> 00:25:57,523
PREVIOUSLY KNOWN TO BE REQUIRED

728
00:25:57,523 --> 00:25:58,724
FOR CMG ASSEMBLY

729
00:25:58,724 --> 00:26:00,292
MCM3, ONE OF THE MCM

730
00:26:00,292 --> 00:26:01,560
TWO THROUGH SEVEN SUBUNITS.

731
00:26:01,560 --> 00:26:01,860
SLD

732
00:26:01,860 --> 00:26:03,762
FIVE ONE OF THE GINS SUBUNITS

733
00:26:03,762 --> 00:26:05,931
AND DNA POLYMERASE EPSILON

734
00:26:05,931 --> 00:26:08,634
THE SECOND SUBUNIT,

735
00:26:08,634 --> 00:26:09,968
WHICH IN YEAST

736
00:26:09,968 --> 00:26:11,370
AT LEAST HAS BEEN IMPLICATED

737
00:26:11,370 --> 00:26:13,138
IN CMG ASSEMBLY.

738
00:26:13,138 --> 00:26:14,306
SO YOU CAN CLICK ON ANY

739
00:26:14,306 --> 00:26:17,042
ONE OF THESE PARTNERS AND RETRIEVE

740
00:26:17,042 --> 00:26:20,045
THE ACTUAL STRUCTURE PREDICTION, WHICH,

741
00:26:20,145 --> 00:26:23,148
HAS ATOMIC RESOLUTION.

742
00:26:23,682 --> 00:26:27,286
AND WHEN WE TOOK ALL OF THESE PREDICTIONS

743
00:26:27,286 --> 00:26:28,854
AND ASSEMBLED THEM

744
00:26:28,854 --> 00:26:31,223
INTO A SINGLE PICTURE, THE

745
00:26:31,223 --> 00:26:33,492
OR A POSSIBLE MECHANISM OF HOW DONSON

746
00:26:33,492 --> 00:26:34,526
AND PROMOTE

747
00:26:34,526 --> 00:26:36,195
CMG ASSEMBLY

748
00:26:36,195 --> 00:26:37,229
REALLY QUITE LITERALLY

749
00:26:37,229 --> 00:26:38,163
LEAPT OFF THE PAGE.

750
00:26:38,163 --> 00:26:41,467
AND THAT WAS THAT DONSON WOULD ORGANIZE

751
00:26:41,934 --> 00:26:43,836
TOPBP1, GINS

752
00:26:43,836 --> 00:26:45,838
AND POL EPSILON INTO WHAT WE NOW

753
00:26:45,838 --> 00:26:47,940
CALL A PRE-LOADING COMPLEX,

754
00:26:47,940 --> 00:26:49,308
AND THAT IT WOULD USE,

755
00:26:49,308 --> 00:26:50,776
ITS PREDICTED INTERACTION

756
00:26:50,776 --> 00:26:54,246
WITH MCM THREE TO REALLY GUIDE THE GINS

757
00:26:54,780 --> 00:26:55,814
HELICASE SUBUNIT

758
00:26:55,814 --> 00:26:59,485
TO ITS PRECISE LOCATION ON THE HELICASE

759
00:26:59,952 --> 00:27:01,987
AND THROUGH,

760
00:27:01,987 --> 00:27:04,523
A LONG SERIES OF STRUCTURE FUNCTION.

761
00:27:04,523 --> 00:27:06,124
EXPERIMENTS

762
00:27:06,124 --> 00:27:08,927
WHERE WE MUTATED THE PREDICTED INTERFACES

763
00:27:08,927 --> 00:27:11,930
AND ASKED WHETHER THE PRELOADING

764
00:27:11,930 --> 00:27:13,131
COMPLEX STILL FORMS

765
00:27:13,131 --> 00:27:15,133
AND WHETHER REPLICATION STILL OCCURS.

766
00:27:15,133 --> 00:27:16,568
WE WERE ABLE TO PROVIDE

767
00:27:16,568 --> 00:27:18,370
COMPELLING SUPPORT FOR THIS MODEL.

768
00:27:18,370 --> 00:27:20,272
AND I SHOULD ALSO MENTION

769
00:27:20,272 --> 00:27:21,440
THAT THIS WAS ONE OF THESE

770
00:27:21,440 --> 00:27:22,875
PERFECT STORM SITUATIONS

771
00:27:22,875 --> 00:27:25,878
WHERE A NUMBER OF OTHER GROUPS ALSO,

772
00:27:26,378 --> 00:27:28,514
WERE ABLE TO IMPLICATE DONSON

773
00:27:28,514 --> 00:27:30,749
IN THE ASSEMBLY OF CMG.

774
00:27:30,749 --> 00:27:33,752
AND SOME OF THESE GROUPS ACTUALLY ALSO,

775
00:27:34,119 --> 00:27:37,122
PROVIDED EXPERIMENTAL STRUCTURE

776
00:27:37,389 --> 00:27:40,726
DETERMINATION THAT AGREED VERY NICELY

777
00:27:40,726 --> 00:27:42,861
WITH, WITH THE STRUCTURE PREDICTION.

778
00:27:42,861 --> 00:27:45,797
SO THIS WAS A REALLY GRATIFYING CASE

779
00:27:45,797 --> 00:27:49,167
WHERE WE WERE ABLE TO USE THIS IN SILICO

780
00:27:49,167 --> 00:27:50,469
SCREENING APPROACH

781
00:27:50,469 --> 00:27:53,472
TO IDENTIFY FUNCTIONALLY RELEVANT

782
00:27:53,472 --> 00:27:56,375
INTERACTORS OF AN OBSCURE PROTEIN,

783
00:27:56,375 --> 00:27:57,342
SUCH AS DONSON.

784
00:27:58,310 --> 00:28:01,313
SO, NOW I'LL TELL YOU ABOUT

785
00:28:02,147 --> 00:28:04,516
SORT OF, I GUESS THE SIMILAR STORY

786
00:28:04,516 --> 00:28:06,251
WHERE WHERE

787
00:28:06,251 --> 00:28:08,353
THIS IN SILICO SCREENING APPROACH

788
00:28:08,353 --> 00:28:09,154
REALLY CRACKED

789
00:28:09,154 --> 00:28:10,088
OPEN THE,

790
00:28:10,088 --> 00:28:13,625
THE QUESTION OF HOW STK19 PROMOTES TC-NER.

791
00:28:13,625 --> 00:28:15,427
SO WHEN WE LOOKED

792
00:28:15,427 --> 00:28:17,729
AT THE LIST,

793
00:28:17,729 --> 00:28:19,531
THE PRIORITIZED LIST

794
00:28:19,531 --> 00:28:23,201
OF PREDICTED STK19 INTERACTORS,

795
00:28:23,201 --> 00:28:25,003
THE REALLY STRIKING FACT

796
00:28:25,003 --> 00:28:27,873
WAS THAT AMONG THE 300

797
00:28:27,873 --> 00:28:29,174
GENOME MAINTENANCE PROTEINS,

798
00:28:29,174 --> 00:28:30,642
THE TOP TWO HITS

799
00:28:30,642 --> 00:28:34,179
WERE THE CSA SUBUNIT OF THE CRL4CSA

800
00:28:34,246 --> 00:28:35,213
LIGASE,

801
00:28:35,213 --> 00:28:38,750
AND THE XPD SUBUNIT OF THE TFIIHIIH COMPLEX.

802
00:28:39,251 --> 00:28:41,119
AND THIS PRETTY MUCH IMMEDIATELY

803
00:28:41,119 --> 00:28:44,022
SUGGESTED A POSSIBLE MECHANISM

804
00:28:44,022 --> 00:28:44,523
BY WHICH

805
00:28:44,523 --> 00:28:46,558
STK19 PROMOTES TC-NER,

806
00:28:46,558 --> 00:28:48,327
AND THAT IS TO INCORPORATE

807
00:28:48,327 --> 00:28:50,562
INTO THE TC-NER COMPLEX

808
00:28:50,562 --> 00:28:51,530
AND THEN USE THE PROTEIN

809
00:28:51,530 --> 00:28:57,035
PROTEIN INTERACTION WITH XPD TO, RECRUIT

810
00:28:57,569 --> 00:29:00,272
OR AT LEAST POSITION TFIIH

811
00:29:00,272 --> 00:29:01,139
IN SUCH A WAY

812
00:29:01,139 --> 00:29:02,441
THAT XPD IS GUIDED

813
00:29:02,441 --> 00:29:04,543
TO THE TEMPLATE STRAND.

814
00:29:04,543 --> 00:29:06,244
SO IN ORDER TO ACTUALLY

815
00:29:06,244 --> 00:29:08,413
TO TEST THE VALIDITY OF THIS MODEL,

816
00:29:08,413 --> 00:29:09,615
WE DID A FEW THINGS.

817
00:29:09,615 --> 00:29:10,515
THE FIRST,

818
00:29:10,515 --> 00:29:13,251
I'LL FOCUS ON THE FIRST PART OF IT,

819
00:29:13,251 --> 00:29:16,455
WHICH IS THE IDEA THAT STK19 FORMS

820
00:29:16,455 --> 00:29:18,290
PART OF THE TC-NER COMPLEX.

821
00:29:18,290 --> 00:29:19,625
AND WE DID TWO THINGS.

822
00:29:19,625 --> 00:29:20,692
THE FIRST

823
00:29:20,692 --> 00:29:21,360
THAT WE COULD

824
00:29:21,360 --> 00:29:23,128
INITIATE IMMEDIATELY, ONCE WE HAD

825
00:29:23,128 --> 00:29:25,163
THE PREDICTION, WAS TO MUTAGENIZE

826
00:29:25,163 --> 00:29:26,632
THE PREDICTED INTERFACE,

827
00:29:26,632 --> 00:29:29,635
GUIDED BY THE ACTUAL MOLECULAR

828
00:29:30,135 --> 00:29:33,672
RESOLUTION STRUCTURE PREDICTION.

829
00:29:34,172 --> 00:29:36,975
AND WHEN WE MADE MUTATIONS IN STK

830
00:29:36,975 --> 00:29:38,076
19 THAT ARE PREDICTED

831
00:29:38,076 --> 00:29:40,245
TO DISRUPT THE INTERACTION WITH CSA,

832
00:29:40,245 --> 00:29:41,813
THE RESULTING STK19

833
00:29:41,813 --> 00:29:43,849
MUTANT WAS TOTALLY INACTIVE

834
00:29:43,849 --> 00:29:45,584
FOR CELL FREE TC-NER.

835
00:29:45,584 --> 00:29:48,553
SO THAT WAS CONSISTENT WITH THE IDEA.

836
00:29:48,553 --> 00:29:49,688
BUT WE ACTUALLY,

837
00:29:49,688 --> 00:29:51,523
WERE REALLY, REALLY LUCKY

838
00:29:51,523 --> 00:29:53,058
THAT WE WERE ABLE TO

839
00:29:53,058 --> 00:29:54,026
THEN COLLABORATE

840
00:29:54,026 --> 00:29:55,761
WITH MY COLLEAGUE LUCAS,

841
00:29:55,761 --> 00:29:58,063
FARNUNG IN THE CELL BIOLOGY DEPARTMENT.

842
00:29:58,063 --> 00:30:01,333
AND, HE HAD PRETTY MUCH EVERYTHING

843
00:30:01,867 --> 00:30:04,636
LYING AROUND, IN HIS FREEZERS

844
00:30:04,636 --> 00:30:05,604
THAT WAS NEEDED

845
00:30:05,604 --> 00:30:08,640
TO ASSEMBLE THE TC-NER COMPLEX.

846
00:30:08,974 --> 00:30:12,911
AND WE GAVE HIM STK19 AND ALL TOGETHER.

847
00:30:12,911 --> 00:30:14,913
THEN HE WAS ABLE TO SOLVE

848
00:30:14,913 --> 00:30:18,884
A, A CRYO-EM STRUCTURE OF AN STK19

849
00:30:18,884 --> 00:30:20,185
CONTAINING TC-NER

850
00:30:20,185 --> 00:30:21,953
COMPLEX AT A REALLY IMPRESSIVE

851
00:30:21,953 --> 00:30:25,223
RESOLUTION OF 1.9 ANGSTROM, AND THAT

852
00:30:25,757 --> 00:30:28,560
VERY, BEAUTIFULLY CONFIRMED

853
00:30:28,560 --> 00:30:31,630
THE INTERACTION OF STK19 WITH CSA.

854
00:30:31,630 --> 00:30:34,633
SO, THAT STRUCTURE LOOKS ALMOST EXACTLY,

855
00:30:34,933 --> 00:30:36,568
AS PREDICTED BY ALPHAFOLD.

856
00:30:36,568 --> 00:30:39,871
BUT IT ALSO SHOWED THAT STK19

857
00:30:39,871 --> 00:30:41,640
CAN BIND SIMULTANEOUSLY

858
00:30:41,640 --> 00:30:43,909
WITH CSA AND RBP1,

859
00:30:43,909 --> 00:30:46,745
AND THAT IT ALSO MAKES CLOSE CONTACT

860
00:30:46,745 --> 00:30:50,282
WITH THE DOWNSTREAM DNA, THAT IS, THE DNA

861
00:30:50,282 --> 00:30:53,919
THAT IS IN THE DIRECTION OF, OF MOVEMENT

862
00:30:54,252 --> 00:30:57,255
BEFORE RNA POLYMERASE GETS STALLED.

863
00:30:57,789 --> 00:31:01,026
SO SO THIS REALLY SUPPORTED THE IDEA

864
00:31:01,026 --> 00:31:02,127
THAT STK19

865
00:31:02,127 --> 00:31:03,095
IS AN INTEGRAL

866
00:31:03,095 --> 00:31:06,098
STRUCTURAL COMPONENT
OF THE TC-NER COMPLEX.

867
00:31:06,264 --> 00:31:07,032
AND SO NOW WE WANTED

868
00:31:07,032 --> 00:31:08,400
TO TEST THE OTHER IDEA,

869
00:31:08,400 --> 00:31:11,570
WHICH IS THAT STK19 RECRUITS TFIIH.

870
00:31:11,803 --> 00:31:12,304
AND AGAIN

871
00:31:12,304 --> 00:31:15,373
WE STARTED WITH A MUTAGENESIS

872
00:31:15,373 --> 00:31:17,709
GUIDED BY THE STRUCTURE PREDICTION.

873
00:31:17,709 --> 00:31:18,076
AND AGAIN

874
00:31:18,076 --> 00:31:19,778
WE FOUND THAT WHEN WE MUTAGENIZE

875
00:31:19,778 --> 00:31:20,812
THIS INTERFACE

876
00:31:20,812 --> 00:31:23,815
THAT THE RESULTING STK19 PROTEIN

877
00:31:23,815 --> 00:31:27,986
WAS LARGELY INACTIVE FOR CELL FREE TC-NER.

878
00:31:29,087 --> 00:31:31,523
NOW, WE WOULD OF COURSE

879
00:31:31,523 --> 00:31:34,392
LOVE TO ALSO SOLVE THE STRUCTURE

880
00:31:34,392 --> 00:31:36,228
OF THIS ENTIRE ASSEMBLY,

881
00:31:36,228 --> 00:31:37,662
AND WE ARE WORKING ON THAT.

882
00:31:37,662 --> 00:31:38,597
BUT IN THE INTERIM,

883
00:31:38,597 --> 00:31:41,399
WHAT WE DID WAS TO ACTUALLY MODEL

884
00:31:41,399 --> 00:31:44,503
HOW THIS, THIS ASSEMBLY MIGHT LOOK.

885
00:31:44,903 --> 00:31:46,004
AND THE WAY THAT WE DID

886
00:31:46,004 --> 00:31:47,572
THAT WAS TO START

887
00:31:47,572 --> 00:31:50,575
WITH THE CRYO-EM STRUCTURE OF THE SDK

888
00:31:50,575 --> 00:31:53,578
19 CONTAINING TC-NER COMPLEX.

889
00:31:53,645 --> 00:31:56,481
AND WE THEN TOOK THE ALPHAFOLD PREDICTION

890
00:31:56,481 --> 00:31:59,718
OF THE STK19 PD INTERACTION.

891
00:31:59,718 --> 00:32:01,419
AND WE ALIGNED IT

892
00:32:01,419 --> 00:32:03,088
TO THE CRYO-EM STRUCTURE

893
00:32:03,088 --> 00:32:05,957
USING THE STK19 SUBUNIT ON THE ONE HAND.

894
00:32:05,957 --> 00:32:07,359
AND THEN ON THE OTHER HAND,

895
00:32:07,359 --> 00:32:10,662
WE TOOK A PREVIOUSLY DETERMINED AGAIN

896
00:32:10,662 --> 00:32:11,830
BY THE KRAMER LAB

897
00:32:11,830 --> 00:32:13,331
CRYO-EM STRUCTURE

898
00:32:13,331 --> 00:32:16,935
OF THE ENTIRE TFIIH COMPLEX IN ASSOCIATION

899
00:32:17,502 --> 00:32:20,372
WITH A SPLAYED DNA

900
00:32:20,372 --> 00:32:23,241
SUBSTRATE THAT SORT OF MIMICS

901
00:32:23,241 --> 00:32:27,012
THE TFIIH COMPLEX IN THE ACT OF LESION

902
00:32:27,012 --> 00:32:27,813
VERIFICATION.

903
00:32:27,813 --> 00:32:30,949
HERE, YOU CAN SEE THE SINGLE STRANDED,

904
00:32:30,949 --> 00:32:33,952
DNA GOING THROUGH THE XPD HELIX,

905
00:32:33,952 --> 00:32:37,455
AND WE ALIGNED THIS COMPLEX ON THE PRIOR

906
00:32:37,455 --> 00:32:40,058
ASSEMBLY USING THE XPD SUBUNIT.

907
00:32:40,058 --> 00:32:42,961
AND IF YOU LOOK AT THIS IN DETAIL,

908
00:32:42,961 --> 00:32:45,964
THE STRIKING FACT IS THAT THE 3’

909
00:32:45,964 --> 00:32:47,632
END OF THE SINGLE STRAND

910
00:32:47,632 --> 00:32:49,034
EMERGING FROM XPD,

911
00:32:49,034 --> 00:32:51,303
THAT IS THE STRAND THAT IS UNDERGOING

912
00:32:51,303 --> 00:32:52,604
LESION VERIFICATION,

913
00:32:52,604 --> 00:32:53,538
OR AT LEAST MIMICKING

914
00:32:53,538 --> 00:32:54,539
THE PROCESS OF LESION

915
00:32:54,539 --> 00:32:57,909
VERIFICATION, IS VERY CLOSELY JUXTAPOSED

916
00:32:58,243 --> 00:32:59,511
WITH THE 5’

917
00:32:59,511 --> 00:33:01,446
END OF THE TEMPLATE STRAND

918
00:33:01,446 --> 00:33:03,215
THAT IS PART OF THE 

919
00:33:03,215 --> 00:33:05,917
THE STALLED, TC-NER COMPLEX.

920
00:33:05,917 --> 00:33:07,986
SO THIS IS REALLY ENTIRELY

921
00:33:07,986 --> 00:33:09,688
CONSISTENT WITH THE IDEA

922
00:33:10,822 --> 00:33:11,489
THAT STK

923
00:33:11,489 --> 00:33:15,026
19 REALLY GUIDES THE TFIIH COMPLEX

924
00:33:15,560 --> 00:33:18,363
TO THE RIGHT POSITION, SO THAT AFTER

925
00:33:18,363 --> 00:33:21,967
XPB UNWINDS THE DUPLEX, XPD

926
00:33:22,200 --> 00:33:24,236
WOULD BE ABLE TO ENGAGE

927
00:33:24,236 --> 00:33:26,304
WITH THE TEMPLATE STRAND.

928
00:33:26,304 --> 00:33:27,706
AND SO IN OTHER WORDS,

929
00:33:27,706 --> 00:33:29,741
WE REALLY THINK NOW THAT STK19

930
00:33:29,741 --> 00:33:30,876
IS THE MISSING LINK

931
00:33:30,876 --> 00:33:32,744
BETWEEN THE TC-NER COMPLEX

932
00:33:32,744 --> 00:33:34,212
AND THE DOWNSTREAM

933
00:33:34,212 --> 00:33:35,714
TFIIH DEPENDENT LESION

934
00:33:35,714 --> 00:33:37,782
VERIFICATION PROCESS.

935
00:33:37,782 --> 00:33:40,952
AND SO IF WE NOW THINK ABOUT THIS

936
00:33:40,952 --> 00:33:43,955
IN TERMS OF, YOU KNOW, THE WHOLE MODEL,

937
00:33:44,022 --> 00:33:45,190
WHAT WE'RE THINKING

938
00:33:45,190 --> 00:33:46,424
IS THAT AFTER RNA POLYMERASE

939
00:33:46,424 --> 00:33:48,226
STALLS AT A LESION,

940
00:33:48,226 --> 00:33:51,463
THERE IS THE ASSEMBLY OF THE TC-NER COMPLEX

941
00:33:51,463 --> 00:33:55,100
THAT WE NOW APPRECIATE INCLUDES STK19.

942
00:33:55,901 --> 00:33:58,737
THE NEXT STEP IS THE RECRUITMENT OF 

943
00:33:58,737 --> 00:33:59,204
TFIIH

944
00:33:59,204 --> 00:34:01,640
AND I DIDN'T MENTION THIS BEFORE,

945
00:34:01,640 --> 00:34:02,874
BUT I'LL MENTION IT NOW.

946
00:34:02,874 --> 00:34:05,911
SEVERAL GROUPS HAVE SHOWN THAT

947
00:34:06,177 --> 00:34:08,413
IN ORDER TO RECRUIT

948
00:34:08,413 --> 00:34:11,016
TFIIH TO A TC-NER COMPLEX,

949
00:34:11,016 --> 00:34:12,050
YOU ACTUALLY REQUIRE

950
00:34:12,050 --> 00:34:13,251
AN INTERACTION

951
00:34:13,251 --> 00:34:16,021
BETWEEN THE P62 SUBUNIT OF TFIIH

952
00:34:16,021 --> 00:34:19,224
AND THIS TFIIH INTERACTING REGION IN,

953
00:34:19,591 --> 00:34:21,192
A SHORT PEPTIDE

954
00:34:21,192 --> 00:34:21,893
THAT SITS

955
00:34:21,893 --> 00:34:22,994
ON AN UNSTRUCTURED

956
00:34:22,994 --> 00:34:25,730
LOOP OF THE UVSSA PROTEIN.

957
00:34:25,730 --> 00:34:29,067
SO WE ACTUALLY THINK THAT UVSSA IS

958
00:34:29,067 --> 00:34:31,870
THE SORT OF INITIAL CAPTURE

959
00:34:31,870 --> 00:34:33,872
POINT FOR TFIIH.

960
00:34:33,872 --> 00:34:36,942
BUT WE BELIEVE THAT THE PROTEIN

961
00:34:36,942 --> 00:34:37,876
PROTEIN INTERACTION

962
00:34:37,876 --> 00:34:41,513
BETWEEN STK19 AND XPD IS WHAT ULTIMATELY

963
00:34:41,746 --> 00:34:44,883
PROPERLY POSITIONS TFIIH IN PRECISELY

964
00:34:45,717 --> 00:34:47,452
THE CONFIGURATION THAT IS NEEDED

965
00:34:47,452 --> 00:34:50,455
FOR PRODUCTIVE LESION VERIFICATION.

966
00:34:50,689 --> 00:34:52,557
AND OF COURSE, THE DOWNSTREAM EVENTS

967
00:34:52,557 --> 00:34:56,094
THEN ARE ASSEMBLY OF AN INCISION

968
00:34:56,094 --> 00:34:59,097
COMPLEX AND GAP FILLING.

969
00:34:59,264 --> 00:35:02,500
NOW, I WOULD LIKE TO, POINT OUT

970
00:35:02,500 --> 00:35:03,902
ONE ASPECT OF THIS

971
00:35:03,902 --> 00:35:05,804
THAT IS DESCRIBED IN DETAIL

972
00:35:05,804 --> 00:35:07,339
IN TYCHO'S PAPER.

973
00:35:07,339 --> 00:35:08,940
I DON'T HAVE TIME TO SHOW YOU THE DATA,

974
00:35:08,940 --> 00:35:10,542
BUT I WANT TO POINT OUT

975
00:35:10,542 --> 00:35:13,578
THAT ACTUALLY THE CRYO-EM STRUCTURE

976
00:35:14,079 --> 00:35:16,915
OF THE STK19 CONTAINING TC

977
00:35:16,915 --> 00:35:20,986
NER COMPLEX REALLY SHOWS THAT,

978
00:35:21,419 --> 00:35:25,957
THE THE VHS DOMAIN OF UVSSA

979
00:35:26,591 --> 00:35:30,462
ACTUALLY BLOCKS THE SITE ON STK19,

980
00:35:30,462 --> 00:35:33,665
WHERE XPD WOULD NORMALLY BE RECRUITED.

981
00:35:33,765 --> 00:35:36,134
AND SO THAT WAS A BIT OF A PARADOX.

982
00:35:36,134 --> 00:35:38,703
BUT THEN WE REALIZED THAT ACTUALLY

983
00:35:38,703 --> 00:35:41,673
UVSSA INTERACTS WITH THIS FLEXIBLE,

984
00:35:41,673 --> 00:35:45,677
C-TERMINAL TAIL OF CSA

985
00:35:46,277 --> 00:35:49,614
THAT WE THINK ALLOWS UVSSA

986
00:35:49,614 --> 00:35:51,416
TO SWING OUT OF THE WAY

987
00:35:51,416 --> 00:35:53,952
TO MAKE ROOM FOR XPD.

988
00:35:53,952 --> 00:35:56,688
AND IN FACT, WE MADE A SINGLE POINT

989
00:35:56,688 --> 00:35:59,691
MUTATION IN THIS C-TERMINAL TAIL OF CSA,

990
00:36:00,091 --> 00:36:02,827
THAT WE SHOWED TOTALLY DISRUPTS THE,

991
00:36:02,827 --> 00:36:05,830
THE ABILITY OF CSA TO SUPPORT,

992
00:36:05,897 --> 00:36:08,800
TC-NER, IN TOTAL AGREEMENT WITH,

993
00:36:08,800 --> 00:36:09,768
WITH THIS MODEL.

994
00:36:10,969 --> 00:36:12,003
SO THAT WAS ONE,

995
00:36:12,003 --> 00:36:13,638
SORT OF ADDITIONAL WRINKLE

996
00:36:13,638 --> 00:36:15,040
THAT I WANTED TO POINT OUT.

997
00:36:15,040 --> 00:36:18,043
NOW, THERE ARE OTHER, SORT OF

998
00:36:18,710 --> 00:36:19,878
OBVIOUS QUESTIONS

999
00:36:19,878 --> 00:36:21,546
THAT THAT RESULT FROM THIS MODEL,

1000
00:36:21,546 --> 00:36:25,550
ONE OF WHICH IS HOW THE SO-CALLED CAK

1001
00:36:25,550 --> 00:36:29,487
DOMAIN OF THE TFIIH COMPLEX IS REMOVED.

1002
00:36:29,487 --> 00:36:31,089
SO IT'S KNOWN THAT THE CAK

1003
00:36:31,089 --> 00:36:32,957
DOMAIN IS REQUIRED FOR TFIIH,

1004
00:36:32,957 --> 00:36:35,293
ITS FUNCTION IN TRANSCRIPTION INITIATION,

1005
00:36:35,293 --> 00:36:37,328
BUT IT'S ACTUALLY INHIBITORY

1006
00:36:37,328 --> 00:36:40,331
FOR TC-NER AND HAS TO BE REMOVED.

1007
00:36:40,732 --> 00:36:43,301
AND, THE QUESTION IS,

1008
00:36:43,301 --> 00:36:46,304
AT WHAT STAGE OF THIS RECRUITMENT

1009
00:36:46,304 --> 00:36:48,706
OR THIS CAPTURE AND DOCKING,

1010
00:36:48,706 --> 00:36:50,408
PATHWAY FOR TFIIH

1011
00:36:50,408 --> 00:36:52,477
 THE CAK MODULE IS REMOVED

1012
00:36:52,477 --> 00:36:53,144
AND THAT'S SOMETHING

1013
00:36:53,144 --> 00:36:56,147
THAT WE ARE ACTIVELY THINKING ABOUT.

1014
00:36:56,414 --> 00:36:57,482
AND THEN OF COURSE,

1015
00:36:57,482 --> 00:36:59,250
YOU KNOW, BACK TO

1016
00:36:59,250 --> 00:37:00,752
SOME OF THE CLASSIC QUESTIONS

1017
00:37:00,752 --> 00:37:02,954
LIKE WHAT DOES RNA POLYMERASE

1018
00:37:02,954 --> 00:37:04,089
ACTUALLY DO?

1019
00:37:04,089 --> 00:37:05,223
WHAT IS ITS FATE

1020
00:37:05,223 --> 00:37:08,326
DURING DURING THESE, VARIOUS,

1021
00:37:08,793 --> 00:37:10,361
REPAIR STEPS?

1022
00:37:10,361 --> 00:37:12,797
DOES IT DOES IT BACKTRACK AND RESTART

1023
00:37:12,797 --> 00:37:14,399
OR IS IT REMOVED?

1024
00:37:14,399 --> 00:37:15,800
THESE ARE ALSO VERY,

1025
00:37:15,800 --> 00:37:17,802
I THINK INTERESTING QUESTIONS THAT

1026
00:37:17,802 --> 00:37:19,737
THAT PERHAPS NOW BECOME TRACTABLE.

1027
00:37:21,439 --> 00:37:22,173
OKAY.

1028
00:37:22,173 --> 00:37:23,208
SO THAT'S

1029
00:37:23,208 --> 00:37:23,741
WHAT I WANTED

1030
00:37:23,741 --> 00:37:26,711
TO TELL YOU ABOUT TC-NER

1031
00:37:27,045 --> 00:37:29,814
AND HOW STRUCTURE-PREDICTION

1032
00:37:29,814 --> 00:37:32,584
REALLY ACCELERATED OUR PROGRESS IN

1033
00:37:32,584 --> 00:37:35,653
IN UNDERSTANDING HOW THIS NEW FACTOR,

1034
00:37:35,653 --> 00:37:38,623
STK19 PROMOTES THIS REACTION.

1035
00:37:38,957 --> 00:37:40,758
BUT WHAT I WANT TO DO IN THE LAST

1036
00:37:40,758 --> 00:37:41,960
COUPLE OF MINUTES

1037
00:37:41,960 --> 00:37:45,130
HERE IS TO ASK THE QUESTION,

1038
00:37:45,130 --> 00:37:45,763
WHAT ARE REALLY

1039
00:37:45,763 --> 00:37:48,399
THE LIMITS OF THIS IN SILICO

1040
00:37:48,399 --> 00:37:50,568
SCREENING APPROACH?

1041
00:37:50,568 --> 00:37:52,370
SO

1042
00:37:52,370 --> 00:37:54,539
IN OTHER WORDS,

1043
00:37:54,539 --> 00:37:56,274
CONSIDER AN EXTREME SCENARIO

1044
00:37:56,274 --> 00:37:58,176
WHERE YOU HAVE A PROTEIN

1045
00:37:58,176 --> 00:38:00,812
THAT, YOU FIND INTERESTING

1046
00:38:00,812 --> 00:38:02,113
FOR WHATEVER REASON.

1047
00:38:02,113 --> 00:38:03,448
BUT LET'S SAY, YOU KNOW

1048
00:38:03,448 --> 00:38:05,650
ESSENTIALLY NOTHING ABOUT IT.

1049
00:38:05,650 --> 00:38:08,319
COULD YOU ACTUALLY CARRY OUT

1050
00:38:08,319 --> 00:38:11,322
A PROTEOME WIDE IN SILICO SCREEN

1051
00:38:11,656 --> 00:38:13,658
IN ORDER TO IDENTIFY

1052
00:38:13,658 --> 00:38:14,792
FUNCTIONALLY RELEVANT

1053
00:38:14,792 --> 00:38:15,827
INTERACTORS, WHICH,

1054
00:38:15,827 --> 00:38:16,728
OF COURSE,

1055
00:38:16,728 --> 00:38:19,664
WOULD PROBABLY TELL YOU A LOT ABOUT

1056
00:38:19,664 --> 00:38:20,865
NOT ONLY THE PATHWAY

1057
00:38:20,865 --> 00:38:22,300
IN WHICH THE PROTEIN OPERATES,

1058
00:38:22,300 --> 00:38:25,303
BUT ALSO ABOUT ITS MECHANISM.

1059
00:38:25,837 --> 00:38:29,841
SO, SO WE WERE VERY CURIOUS ABOUT THIS.

1060
00:38:30,341 --> 00:38:34,045
SO WE WENT AHEAD AND ACTUALLY SCALED UP

1061
00:38:34,145 --> 00:38:35,280
SOME OF THE SCREENS

1062
00:38:35,280 --> 00:38:36,214
THAT I'VE TOLD YOU ABOUT.

1063
00:38:36,214 --> 00:38:39,517
SO, WE WENT AHEAD AND FOLDED STK19

1064
00:38:39,517 --> 00:38:42,720
WITH ALL 20,000 HUMAN PROTEINS.

1065
00:38:42,720 --> 00:38:45,723
AND THEN WE USED VARIOUS METRICS TO ASK,

1066
00:38:46,057 --> 00:38:47,525
WHERE

1067
00:38:47,525 --> 00:38:49,827
IN THIS LIST OF 20,000

1068
00:38:49,827 --> 00:38:51,663
BINARY STRUCTURE PREDICTIONS,

1069
00:38:51,663 --> 00:38:54,532
WOULD THE FUNCTIONALLY RELEVANT PARTNERS

1070
00:38:54,532 --> 00:38:56,301
THAT WE NOW,

1071
00:38:56,301 --> 00:38:57,569
KNOW ABOUT,

1072
00:38:57,569 --> 00:39:00,471
WHERE WOULD THEY SIT IN THIS RANK LIST?

1073
00:39:00,471 --> 00:39:02,173
AND SO IF YOU TAKE ONE OF THE,

1074
00:39:02,173 --> 00:39:05,176
VERY POPULAR CONFIDENCE METRICS

1075
00:39:05,176 --> 00:39:06,277
CALLED PDOCKQ,

1076
00:39:06,277 --> 00:39:08,413
THE RESULTS ARE REALLY QUITE TERRIBLE.

1077
00:39:08,413 --> 00:39:09,147
SO THE,

1078
00:39:09,147 --> 00:39:12,650
THE FUNCTIONALLY RELEVANT PARTNERS, CSA,

1079
00:39:12,650 --> 00:39:14,986
WHICH ALSO GOES BY THE NAME OF ERCC8

1080
00:39:14,986 --> 00:39:15,820
AND XPD,

1081
00:39:15,820 --> 00:39:18,489
WHICH GOES BY THE NAME OF ERCC2,

1082
00:39:18,489 --> 00:39:20,992
AS WELL AS RNA POLYMERASE

1083
00:39:20,992 --> 00:39:23,561
ARE, YOU KNOW, NOWHERE NEAR THE TOP.

1084
00:39:23,561 --> 00:39:25,296
OKAY. SO THAT LOOKS PRETTY BAD.

1085
00:39:25,296 --> 00:39:27,699
IPTM DOES A LOT BETTER.

1086
00:39:27,699 --> 00:39:29,567
SO NOW THESE THREE FUNCTIONALLY

1087
00:39:29,567 --> 00:39:30,034
RELEVANT

1088
00:39:30,034 --> 00:39:34,706
PARTNERS ARE IN THE TOP 60 OR SO, PAIRS.

1089
00:39:34,706 --> 00:39:36,741
SO THAT'S REALLY QUITE GOOD.

1090
00:39:36,741 --> 00:39:40,578
BUT, YOU KNOW, STILL MAYBE

1091
00:39:40,578 --> 00:39:41,579
NOT GOOD ENOUGH

1092
00:39:41,579 --> 00:39:42,847
IF YOU DIDN'T KNOW ANYTHING

1093
00:39:42,847 --> 00:39:44,449
ABOUT THE PROTEIN.

1094
00:39:44,449 --> 00:39:46,050
WE DID ANOTHER EXAMPLE.

1095
00:39:46,050 --> 00:39:49,020
SO REMEMBER THE DONSON PROTEIN

1096
00:39:49,020 --> 00:39:49,787
THAT IS REQUIRED

1097
00:39:49,787 --> 00:39:51,389
FOR CMG HELICASE

1098
00:39:51,389 --> 00:39:52,790
ASSEMBLY AND THAT INTERACTS

1099
00:39:52,790 --> 00:39:53,791
FUNCTIONALLY

1100
00:39:53,791 --> 00:39:56,794
WITH, THESE 4 OR 5 FACTORS HERE.

1101
00:39:57,061 --> 00:39:58,830
IF YOU USE PDOCKQ,

1102
00:39:58,830 --> 00:40:00,698
THE RESULTS ARE REALLY PRETTY TERRIBLE.

1103
00:40:00,698 --> 00:40:02,267
IF YOU USE IPTM, IT GETS BETTER.

1104
00:40:02,267 --> 00:40:03,601
BUT YOU KNOW, THE

1105
00:40:03,601 --> 00:40:04,636
THE FIVE PARTNERS

1106
00:40:04,636 --> 00:40:07,639
HERE ARE SPREAD OVER THE TOP 1000 HITS.

1107
00:40:07,739 --> 00:40:09,841
AND AND SO AGAIN,

1108
00:40:09,841 --> 00:40:11,576
THAT'S NOT TERRIBLY USEFUL

1109
00:40:11,576 --> 00:40:14,112
IF YOU DON'T KNOW MUCH ABOUT THE PROTEIN

1110
00:40:14,112 --> 00:40:15,413
TO BEGIN WITH.

1111
00:40:15,413 --> 00:40:16,781
SO FOR THIS FOR US,

1112
00:40:16,781 --> 00:40:19,784
THIS REALLY SUGGESTED THAT WE NEED BETTER

1113
00:40:20,118 --> 00:40:21,919
CONFIDENCE METRICS

1114
00:40:21,919 --> 00:40:22,954
TO IDENTIFY

1115
00:40:22,954 --> 00:40:24,722
FUNCTIONALLY RELEVANT INTERACTORS.

1116
00:40:24,722 --> 00:40:29,060
SO WHAT WE DID WAS TO ACTUALLY USE

1117
00:40:29,294 --> 00:40:29,927
SORT OF

1118
00:40:29,927 --> 00:40:31,429
CLASSICAL MACHINE

1119
00:40:31,429 --> 00:40:34,432
LEARNING APPROACHES TO TRAIN A CLASSIFIER

1120
00:40:34,699 --> 00:40:37,368
THAT WE HOPED WOULD BE ABLE TO TELL

1121
00:40:37,368 --> 00:40:39,337
THE DIFFERENCE BETWEEN

1122
00:40:40,571 --> 00:40:41,639
FUNCTIONAL

1123
00:40:41,639 --> 00:40:42,974
ALPHAFOLD PREDICTIONS

1124
00:40:42,974 --> 00:40:45,510
AND SPURIOUS ALPHAFOLD PREDICTIONS.

1125
00:40:45,510 --> 00:40:46,911
AND SO WE USED,

1126
00:40:46,911 --> 00:40:47,445
YOU KNOW, THE

1127
00:40:47,445 --> 00:40:49,347
CLASSIC, APPROACH,

1128
00:40:49,347 --> 00:40:52,350
WHICH IS TO ACTUALLY TRAIN A CLASSIFIER

1129
00:40:52,517 --> 00:40:56,120
ON A A SET OF KNOWN DATA POINTS.

1130
00:40:56,120 --> 00:41:00,024
SO WE TOOK A LARGE NUMBER OF ALPHAFOLD

1131
00:41:00,024 --> 00:41:02,226
BINARY ALPHAFOLD PREDICTIONS,

1132
00:41:02,226 --> 00:41:02,660
THE VAST

1133
00:41:02,660 --> 00:41:05,763
MAJORITY OF WHICH WE KNEW TO BE NEGATIVE

1134
00:41:05,763 --> 00:41:07,031
BECAUSE THEY CORRESPONDED

1135
00:41:07,031 --> 00:41:10,401
TO COMPLETELY RANDOM PAIRINGS,

1136
00:41:10,735 --> 00:41:14,772
AS WELL AS A SET OF PAIRS.

1137
00:41:14,772 --> 00:41:16,507
ALPHAFOLD BINARY

1138
00:41:16,507 --> 00:41:17,508
STRUCTURE PREDICTIONS

1139
00:41:17,508 --> 00:41:19,844
THAT WE PRESUMED TO BE POSITIVE

1140
00:41:19,844 --> 00:41:21,979
BECAUSE THEY CORRESPOND TO PROTEIN PAIRS

1141
00:41:21,979 --> 00:41:22,914
THAT ACTUALLY CROSS

1142
00:41:22,914 --> 00:41:24,282
LINK IN VIVO

1143
00:41:24,282 --> 00:41:27,185
IN CROSSLINKING MASS SPEC EXPERIMENTS.

1144
00:41:27,185 --> 00:41:29,153
SO WE MIXED THESE TOGETHER,

1145
00:41:29,153 --> 00:41:33,658
AND THEN WE TRY TO TRAIN A CLASSIFIER

1146
00:41:34,092 --> 00:41:36,694
THAT IS INFORMED BY,

1147
00:41:36,694 --> 00:41:38,463
A NUMBER OF DIFFERENT FEATURES.

1148
00:41:38,463 --> 00:41:39,630
AND SO THESE FEATURES

1149
00:41:39,630 --> 00:41:42,166
THAT WE FED THE CLASSIFIER FALL

1150
00:41:42,166 --> 00:41:43,534
INTO TWO DIFFERENT GROUPS.

1151
00:41:43,534 --> 00:41:45,903
ONE WE CALL STRUCTURAL FEATURES.

1152
00:41:47,071 --> 00:41:47,538
SO THESE

1153
00:41:47,538 --> 00:41:50,541
ARE NOT ONLY THE

1154
00:41:50,541 --> 00:41:53,978
ALPHAFOLD DERIVED CONFIDENCE METRICS

1155
00:41:53,978 --> 00:41:55,613
THAT COME WITH EVERY BINARY

1156
00:41:55,613 --> 00:41:59,250
STRUCTURE PREDICTION IPTM, PIDDT

1157
00:41:59,484 --> 00:42:00,318
AND SO FORTH,

1158
00:42:00,318 --> 00:42:04,055
BUT ALSO MEASURABLE PROPERTIES OF THE

1159
00:42:04,222 --> 00:42:05,590
PREDICTIONS THEMSELVES.

1160
00:42:05,590 --> 00:42:09,427
SO HOW MANY HYDROGEN AND SALT BRIDGES,

1161
00:42:09,794 --> 00:42:11,629
ARE AT THE INTERFACE?

1162
00:42:11,629 --> 00:42:13,898
WHAT IS THE SIZE OF THE,

1163
00:42:13,898 --> 00:42:16,901
OF THE INTERFACE AND SO FORTH?

1164
00:42:17,068 --> 00:42:18,403
AND WE ACTUALLY INITIALLY

1165
00:42:18,403 --> 00:42:19,737
TRAINED A CLASSIFIER

1166
00:42:19,737 --> 00:42:21,572
USING ONLY THESE STRUCTURAL FEATURES.

1167
00:42:21,572 --> 00:42:23,007
AND ITS PERFORMANCE

1168
00:42:23,007 --> 00:42:24,375
WAS WAS DECENT,

1169
00:42:24,375 --> 00:42:27,378
BUT NOT QUITE WHERE WE WANTED IT TO BE.

1170
00:42:27,612 --> 00:42:28,946
SO WE THEN REASONED THAT

1171
00:42:28,946 --> 00:42:31,015
PERHAPS WE ALSO NEED TO BE LOOKING

1172
00:42:31,015 --> 00:42:32,784
AT THE BIOLOGICAL PROPERTIES

1173
00:42:32,784 --> 00:42:35,353
OF ALL OF THE PROTEIN.

1174
00:42:35,353 --> 00:42:37,321
PAIRS THAT WE'RE LOOKING AT.

1175
00:42:37,321 --> 00:42:41,225
AND SO WE THEN ALSO CONSIDERED,

1176
00:42:41,826 --> 00:42:45,396
OMICS DERIVED BIOLOGICAL FEATURES

1177
00:42:45,396 --> 00:42:48,433
SUCH AS CO-EXPRESSION DATA, CO-LOCALIZATION.

1178
00:42:48,433 --> 00:42:50,902
ARE THE TWO PROTEINS AND THE PAIR,

1179
00:42:50,902 --> 00:42:52,970
DO THEY SHOW A SIMILAR DEPENDENCY

1180
00:42:52,970 --> 00:42:53,738
AND THE THOUSAND

1181
00:42:53,738 --> 00:42:54,505
OR SO

1182
00:42:54,505 --> 00:42:56,007
GENOME WIDE CRISPR SCREENS

1183
00:42:56,007 --> 00:42:57,942
THAT HAVE BEEN REPORTED?

1184
00:42:57,942 --> 00:42:59,544
IS THERE EVIDENCE

1185
00:42:59,544 --> 00:43:00,511
FOR ASSOCIATE

1186
00:43:00,511 --> 00:43:03,314
ASSOCIATION IN BIOGRID, ETC.?

1187
00:43:03,314 --> 00:43:06,584
AND WHEN WE USED BOTH THESE

1188
00:43:06,584 --> 00:43:09,120
STRUCTURAL AND BIOLOGICAL FEATURES

1189
00:43:09,120 --> 00:43:12,423
TO INFORM, AND TRAIN A,

1190
00:43:12,490 --> 00:43:15,493
A CLASSIFIER, A RANDOM FOREST CLASSIFIER

1191
00:43:16,060 --> 00:43:17,929
TO BEST SEPARATE,

1192
00:43:17,929 --> 00:43:20,932
THE POSITIVES AND THE NEGATIVES,

1193
00:43:21,132 --> 00:43:23,868
WE ACTUALLY ENDED UP WITH A CLASSIFIER

1194
00:43:23,868 --> 00:43:25,269
THAT WAS REALLY QUITE USEFUL.

1195
00:43:25,269 --> 00:43:27,238
SO, IN THESE MACHINE

1196
00:43:27,238 --> 00:43:28,539
LEARNING APPROACHES,

1197
00:43:28,539 --> 00:43:32,043
ONE TYPICALLY TAKES 75% OF THE,

1198
00:43:32,343 --> 00:43:35,613
OF THE, OF THE DATA, THE, THE KNOWN

1199
00:43:35,980 --> 00:43:37,815
DATA POINTS TO TRAIN.

1200
00:43:37,815 --> 00:43:40,418
AND THEN ONE USES THE REMAINING 25%

1201
00:43:40,418 --> 00:43:42,253
TO ACTUALLY TEST PERFORMANCE.

1202
00:43:43,321 --> 00:43:44,889
AND, I WILL

1203
00:43:44,889 --> 00:43:48,025
JUST VERY BRIEFLY TELL YOU

1204
00:43:48,092 --> 00:43:51,762
THAT, GUIDED IN LARGE PART

1205
00:43:51,762 --> 00:43:54,298
BY SOME VERY CONSCIENTIOUS

1206
00:43:54,298 --> 00:43:58,102
AND VERY RIGOROUS, COMPUTATIONAL

1207
00:43:58,469 --> 00:44:00,872
BIOLOGY REVIEWERS OF OUR MANUSCRIPT,

1208
00:44:00,872 --> 00:44:02,106
WE PUT OUR CLASSIFIER

1209
00:44:02,106 --> 00:44:03,774
THROUGH ALL OF THE PACES,

1210
00:44:03,774 --> 00:44:04,675
AND,

1211
00:44:04,675 --> 00:44:05,710
AND DID

1212
00:44:05,710 --> 00:44:08,880
A LOT OF THE PERFORMANCE TESTING,

1213
00:44:08,880 --> 00:44:09,881
INCLUDING,

1214
00:44:09,881 --> 00:44:10,515
RUNNING

1215
00:44:10,515 --> 00:44:11,415
THESE ROC CURVES

1216
00:44:11,415 --> 00:44:13,618
AND MEASURING THE AREA UNDER THE CURVE,

1217
00:44:13,618 --> 00:44:16,654
WHICH, I HOPE YOU CAN SEE HERE,

1218
00:44:16,654 --> 00:44:18,523
SHOWS THAT OUR CLASSIFIER,

1219
00:44:18,523 --> 00:44:20,591
WHICH WE CALL SPOC,

1220
00:44:20,591 --> 00:44:21,859
PERFORMS,

1221
00:44:21,859 --> 00:44:22,994
REALLY QUITE WELL

1222
00:44:22,994 --> 00:44:24,629
AND MUCH BETTER THAN

1223
00:44:24,629 --> 00:44:26,964
THE PREVIOUS CONFIDENCE METRICS.

1224
00:44:26,964 --> 00:44:29,133
BUT I'LL SHOW YOU A MUCH MORE INTUITIVE

1225
00:44:29,133 --> 00:44:30,535
PERFORMANCE TEST IN A MOMENT.

1226
00:44:30,535 --> 00:44:34,605
BUT I WILL POINT OUT, THAT, WE CALL IT

1227
00:44:34,605 --> 00:44:35,573
WE CALL IT SPOC,

1228
00:44:35,573 --> 00:44:37,842
NOT ONLY BECAUSE IT'S A CUTE ACRONYM,

1229
00:44:37,842 --> 00:44:39,477
BUT BECAUSE REALLY

1230
00:44:39,477 --> 00:44:42,113
THE CLASSIFIER TAKES INTO CONSIDERATION

1231
00:44:42,113 --> 00:44:44,382
NOT ONLY THE STRUCTURAL FEATURES

1232
00:44:44,382 --> 00:44:45,616
OF THE PROTEIN PAIRS,

1233
00:44:45,616 --> 00:44:49,921
BUT ALSO, A LONG LIST OF OMICS

1234
00:44:50,087 --> 00:44:53,424
DERIVED BIOLOGICAL FEATURES OF EACH PAIR.

1235
00:44:53,424 --> 00:44:55,760
AND IT'S REALLY BY COMBINING,

1236
00:44:55,760 --> 00:44:59,130
BOTH OF THESE, THESE PROPERTIES.

1237
00:44:59,130 --> 00:44:59,864
SO CONSIDERING

1238
00:44:59,864 --> 00:45:02,433
NOT ONLY THE STRUCTURAL PLAUSIBILITY

1239
00:45:02,433 --> 00:45:05,436
OF THE STRUCTURE PREDICTION,

1240
00:45:05,570 --> 00:45:08,239
BUT ALSO THE BIOLOGICAL PROPERTIES

1241
00:45:08,239 --> 00:45:09,907
OF THE PROTEINS INVOLVED,

1242
00:45:09,907 --> 00:45:12,176
THAT WE WERE ABLE TO GENERATE

1243
00:45:12,176 --> 00:45:16,013
A CLASSIFIER, THAT PERFORMS REALLY WELL,

1244
00:45:16,013 --> 00:45:17,682
AND TO CONVINCE YOU THAT IT PERFORMS

1245
00:45:17,682 --> 00:45:19,383
WELL, I WILL REVISIT

1246
00:45:19,383 --> 00:45:20,885
THESE PROTEOME WIDE SCREENS

1247
00:45:20,885 --> 00:45:23,788
AND SHOW YOU THAT UNLIKE THE,

1248
00:45:23,788 --> 00:45:26,824
CLASSIC CONFIDENCE METRICS PDOCKQ AND IPTM,

1249
00:45:26,958 --> 00:45:28,659
WHICH REALLY SPREAD

1250
00:45:28,659 --> 00:45:30,861
THE FUNCTIONALLY RELEVANT PARTNERS

1251
00:45:30,861 --> 00:45:32,630
OVER A LARGE,

1252
00:45:32,630 --> 00:45:35,633
LIST OF PRESUMABLY FALSE POSITIVES,

1253
00:45:35,833 --> 00:45:36,434
SPARK

1254
00:45:36,434 --> 00:45:38,502
REALLY DOES AN AMAZING JOB OF PUTTING

1255
00:45:38,502 --> 00:45:40,371
THE FUNCTIONALLY RELEVANT PARTNERS

1256
00:45:40,371 --> 00:45:41,806
AT THE VERY TOP

1257
00:45:41,806 --> 00:45:44,809
OF THIS LIST OF 20,000 BINARY.

1258
00:45:44,942 --> 00:45:45,610
PREDICTIONS.

1259
00:45:45,610 --> 00:45:48,579
WE SEE THE SAME THING FOR DONSON.

1260
00:45:48,579 --> 00:45:51,983
SO HERE ARE THE FIVE OR SO, FUNCTIONAL.

1261
00:45:51,983 --> 00:45:52,483
RELEVANT

1262
00:45:52,483 --> 00:45:53,851
PARTNERS ARE SPREAD OVER

1263
00:45:53,851 --> 00:45:54,552
THE TOP

1264
00:45:54,552 --> 00:45:57,455
SEVEN HITS IN THIS PROTEOME WIDE SCREEN.

1265
00:45:57,455 --> 00:45:59,423
AND WE'VE ACTUALLY REPEATED,

1266
00:45:59,423 --> 00:46:01,559
VERSIONS OF THESE PROTEOME

1267
00:46:01,559 --> 00:46:04,595
WIDE SCREENS OVER, WE'VE DONE THIS

1268
00:46:04,762 --> 00:46:08,399
ABOUT 30 TIMES AND SPOC CONSISTENTLY

1269
00:46:08,699 --> 00:46:10,001
PERFORMS REALLY WELL,

1270
00:46:10,001 --> 00:46:11,836
INCLUDING ON PROTEINS

1271
00:46:11,836 --> 00:46:13,604
THAT ARE NOT IN THE DNA DAMAGE

1272
00:46:13,604 --> 00:46:16,941
RESPONSE, BUT, OPERATE IN THE CYTOPLASM,

1273
00:46:17,141 --> 00:46:18,542
AND VARIOUS DIFFERENT ORGANELLES.

1274
00:46:18,542 --> 00:46:19,810
AND THAT ACTUALLY MAKES SENSE

1275
00:46:19,810 --> 00:46:21,679
BECAUSE NOTHING ABOUT SPOC'S

1276
00:46:21,679 --> 00:46:25,516
TRAINING WAS ACTUALLY, REALLY PATHWAY

1277
00:46:25,516 --> 00:46:26,150
SPECIFIC.

1278
00:46:27,652 --> 00:46:28,185
OKAY.

1279
00:46:28,185 --> 00:46:31,789
SO NOW THAT WE HAVE A CLASSIFIER

1280
00:46:31,789 --> 00:46:34,291
THAT CAN ACTUALLY PICK OUT

1281
00:46:34,291 --> 00:46:37,061
FUNCTIONALLY RELEVANT

1282
00:46:37,061 --> 00:46:38,896
STRUCTURE PREDICTIONS

1283
00:46:38,896 --> 00:46:42,033
OUT OF A TRULY PROTEOME WIDE SCREEN,

1284
00:46:42,833 --> 00:46:43,601
IN PRINCIPLE,

1285
00:46:43,601 --> 00:46:44,635
WE'RE NOW IN A POSITION

1286
00:46:44,635 --> 00:46:47,538
TO ACTUALLY TRY TO GENERATE

1287
00:46:47,538 --> 00:46:50,875
A COMPREHENSIVE STRUCTURAL INTERACTOME.

1288
00:46:51,242 --> 00:46:53,444
AND SO THAT WOULD REALLY INVOLVE

1289
00:46:53,444 --> 00:46:56,414
TAKING ALL 20,000 HUMAN PROTEINS

1290
00:46:56,580 --> 00:46:59,583
AND PAIRING THEM UP WITH EACH OTHER,

1291
00:46:59,850 --> 00:47:02,753
FOLDING EACH PAIR.

1292
00:47:02,753 --> 00:47:04,522
AND THEN

1293
00:47:04,522 --> 00:47:07,324
APPLYING THE SPOC CLASSIFIER

1294
00:47:07,324 --> 00:47:08,659
THAT ACTUALLY WOULD INVOLVE

1295
00:47:08,659 --> 00:47:11,662
200 MILLION POSSIBLE PAIRS.

1296
00:47:12,463 --> 00:47:14,365
SOMETHING LIKE 0.1% OF

1297
00:47:14,365 --> 00:47:16,434
THESE ARE THOUGHT TO BE,

1298
00:47:16,434 --> 00:47:18,602
ACTUALLY FUNCTIONAL.

1299
00:47:18,602 --> 00:47:20,104
SO IT'S IT'S REALLY LOOKING

1300
00:47:20,104 --> 00:47:21,372
FOR A NEEDLE IN A HAYSTACK,

1301
00:47:21,372 --> 00:47:24,675
WHICH WE THINK, SPOC IS EQUIPPED TO DO.

1302
00:47:24,975 --> 00:47:27,144
HOWEVER, THERE IS OBVIOUSLY ONE

1303
00:47:27,144 --> 00:47:28,012
MAJOR BARRIER,

1304
00:47:28,012 --> 00:47:29,080
AND THAT IS THE COMPUTE

1305
00:47:29,080 --> 00:47:31,348
THAT WOULD BE NEEDED TO

1306
00:47:31,348 --> 00:47:34,785
TO COMPUTE THIS ENTIRE MATRIX.

1307
00:47:35,052 --> 00:47:36,253
SO THIS WOULD COST

1308
00:47:36,253 --> 00:47:37,788
TENS OF MILLIONS OF DOLLARS

1309
00:47:37,788 --> 00:47:40,791
IN, IN CLOUD COMPUTE COMPUTING.

1310
00:47:40,891 --> 00:47:43,694
AND SO WHAT WE AND OTHERS

1311
00:47:43,694 --> 00:47:47,131
ARE DOING TO MAKE THIS SORT

1312
00:47:47,131 --> 00:47:47,865
OF AN APPROACHABLE

1313
00:47:47,865 --> 00:47:52,002
PROBLEM IS TO INTRODUCE A TRIAGE STEP.

1314
00:47:52,002 --> 00:47:53,437
AND I'M HAPPY TO DISCUSS

1315
00:47:53,437 --> 00:47:54,839
HOW WE'RE DOING THIS.

1316
00:47:54,839 --> 00:47:57,842
BUT THE THE IDEA IS TO USE

1317
00:47:58,275 --> 00:47:59,977
MUCH LESS COMPUTATIONALLY

1318
00:47:59,977 --> 00:48:01,345
INTENSIVE APPROACHES

1319
00:48:01,345 --> 00:48:05,616
TO NOMINATE A SMALL SUBSET OF THE 200

1320
00:48:05,616 --> 00:48:07,017
MILLION PAIRS,

1321
00:48:07,017 --> 00:48:09,887
LET'S SAY THE TOP 1% OR 2 MILLION,

1322
00:48:09,887 --> 00:48:10,821
AND THEN SUBJECT

1323
00:48:10,821 --> 00:48:14,458
THOSE TO A FULL ANALYSIS WITH ALPHAFOLD

1324
00:48:14,725 --> 00:48:17,895
AND THE, THE SPOC CLASSIFIER.

1325
00:48:18,429 --> 00:48:20,131
AND WE'RE ACTUALLY GETTING,

1326
00:48:20,131 --> 00:48:21,699
VERY GENEROUS SUPPORT

1327
00:48:21,699 --> 00:48:24,702
FROM NVIDIA CORPORATION TO,

1328
00:48:24,702 --> 00:48:27,671
TO FOLD ABOUT 2 MILLION PAIRS,

1329
00:48:27,772 --> 00:48:31,275
WHICH WE HOPE WILL LEAD TO AT LEAST

1330
00:48:31,275 --> 00:48:33,344
A FIRST DRAFT

1331
00:48:33,344 --> 00:48:36,514
OF A REASONABLE STRUCTURAL INTERACTOME

1332
00:48:36,514 --> 00:48:37,715
THAT WE THINK,

1333
00:48:37,715 --> 00:48:39,784
TOGETHER WITH THE EFFORTS OF OTHER LABS

1334
00:48:39,784 --> 00:48:42,486
LIKE THE CONG LAB AT UT SOUTHWESTERN,

1335
00:48:42,486 --> 00:48:45,589
WILL, YOU KNOW, BEGIN TO GENERATE

1336
00:48:45,589 --> 00:48:48,192
A USEFUL STRUCTURAL INTERACTOME

1337
00:48:48,192 --> 00:48:49,126
THAT CAN CATALYZE

1338
00:48:49,126 --> 00:48:50,628
A LOT OF MECHANISTIC

1339
00:48:50,628 --> 00:48:51,996
DISCOVERY IN THE FIELD.

1340
00:48:53,030 --> 00:48:55,866
SO, UNTIL

1341
00:48:55,866 --> 00:48:58,869
WE HAVE THAT AND AND THAT'S, YOU KNOW,

1342
00:48:59,370 --> 00:49:01,572
GOING TO BE A LITTLE BIT OF,

1343
00:49:01,572 --> 00:49:03,140
IT'S GOING TO TAKE SOME TIME TO,

1344
00:49:03,140 --> 00:49:04,475
TO COMPLETE THE FOLDING.

1345
00:49:04,475 --> 00:49:06,577
WE'RE ABOUT IN THE MIDDLE OF THE RUN

1346
00:49:06,577 --> 00:49:07,511
RIGHT NOW.

1347
00:49:07,511 --> 00:49:09,413
AND IT'S GOING TO REQUIRE

1348
00:49:09,413 --> 00:49:10,781
QUITE A BIT OF ANALYSIS.

1349
00:49:10,781 --> 00:49:11,582
BUT IN THE MEANTIME,

1350
00:49:11,582 --> 00:49:13,150
THERE ARE ACTUALLY RESOURCES

1351
00:49:13,150 --> 00:49:16,987
THAT THAT YOU CAN UTILIZE, ALREADY.

1352
00:49:16,987 --> 00:49:20,191
SO THE PREDICTIVE, SORT RESOURCE,

1353
00:49:20,191 --> 00:49:22,593
WHICH IS REALLY DEDICATED TO,

1354
00:49:22,593 --> 00:49:24,995
THE GENOME MAINTENANCE PATHWAY.

1355
00:49:24,995 --> 00:49:26,463
AND I'D ALSO LIKE TO POINT OUT

1356
00:49:26,463 --> 00:49:27,998
THAT THE CONG LAB AT UT

1357
00:49:27,998 --> 00:49:32,102
SOUTHWESTERN HAS ACTUALLY ALREADY,

1358
00:49:32,102 --> 00:49:33,504
TAKEN A FIRST SWING

1359
00:49:33,504 --> 00:49:37,541
AT A STRUCTURAL INTERACTOME AND,

1360
00:49:37,541 --> 00:49:38,843
PUBLISHED A BIORXIV PAPER,

1361
00:49:38,843 --> 00:49:41,078
AND THEY HAVE THIS VERY NICE WEB RESOURCE

1362
00:49:41,078 --> 00:49:42,112
WHERE YOU CAN TYPE

1363
00:49:42,112 --> 00:49:43,714
IN YOUR FAVORITE PROTEIN,

1364
00:49:43,714 --> 00:49:47,284
AND THEY HAVE NOMINATED ABOUT 8000 NEW,

1365
00:49:47,585 --> 00:49:49,086
VERY HIGH QUALITY, HIGH

1366
00:49:49,086 --> 00:49:51,155
CONFIDENCE PROTEIN-PROTEIN INTERACTIONS.

1367
00:49:51,155 --> 00:49:52,389
WE'VE LOOKED AT THE DATABASE.

1368
00:49:52,389 --> 00:49:54,892
IT'S IT'S IT'S EXCELLENT.

1369
00:49:54,892 --> 00:49:58,195
BUT STILL OF COURSE, VERY INCOMPLETE.

1370
00:49:58,195 --> 00:50:00,164
BUT THIS IS DEFINITELY A STARTING POINT

1371
00:50:00,164 --> 00:50:02,867
FOR ALL OF YOU TO START LOOKING FOR

1372
00:50:02,867 --> 00:50:04,001
FUNCTIONALLY RELEVANT

1373
00:50:04,001 --> 00:50:06,103
INTERACTORS OF YOUR FAVORITE PROTEINS.

1374
00:50:06,103 --> 00:50:06,570
OF COURSE,

1375
00:50:06,570 --> 00:50:07,571
YOU CAN DO

1376
00:50:07,571 --> 00:50:10,674
YOUR OWN PROTEIN FOLDING EXPERIMENTS.

1377
00:50:10,941 --> 00:50:13,644
ESPECIALLY NOW THAT DEEPMIND

1378
00:50:13,644 --> 00:50:14,912
HAS MADE AVAILABLE,

1379
00:50:14,912 --> 00:50:17,214
THIS VERY USER FRIENDLY,

1380
00:50:17,214 --> 00:50:20,251
WEB INTERFACE WHERE YOU CAN,

1381
00:50:20,651 --> 00:50:22,820
USE THE ALPHAFOLD ALGORITHM

1382
00:50:22,820 --> 00:50:24,255
TO LOOK NOT ONLY FOR PROTEIN

1383
00:50:24,255 --> 00:50:25,222
PROTEIN INTERACTIONS,

1384
00:50:25,222 --> 00:50:26,891
BUT ALSO PROTEIN NUCLEIC ACID

1385
00:50:26,891 --> 00:50:28,826
AND PROTEIN LIGAND INTERACTIONS.

1386
00:50:28,826 --> 00:50:30,327
AND IF YOU HAVEN'T USED THIS ALREADY,

1387
00:50:30,327 --> 00:50:32,029
I STRONGLY ENCOURAGE YOU.

1388
00:50:32,029 --> 00:50:32,863
IT'S A BIT

1389
00:50:32,863 --> 00:50:36,267
LIMITED IN HOW MANY, EXPERIMENTS

1390
00:50:36,267 --> 00:50:37,468
YOU CAN DO, BUT YOU CAN

1391
00:50:37,468 --> 00:50:38,569
ACTUALLY, YOU KNOW,

1392
00:50:38,569 --> 00:50:39,003
IF YOU

1393
00:50:39,003 --> 00:50:39,570
IF YOU COLLECT

1394
00:50:39,570 --> 00:50:41,672
A COUPLE OF GOOGLE, ACCOUNTS

1395
00:50:41,672 --> 00:50:42,973
AND DO IT OVER A FEW DAYS,

1396
00:50:42,973 --> 00:50:44,541
YOU CAN EASILY FOLD A FEW HUNDRED

1397
00:50:44,541 --> 00:50:46,043
PAIRS AND DO A MINI SCREEN.

1398
00:50:47,144 --> 00:50:49,079
AND, YOU KNOW, IF THE WHOLE,

1399
00:50:49,079 --> 00:50:50,581
ALPHAFOLD BUSINESS

1400
00:50:50,581 --> 00:50:52,583
IS INTIMIDATING FOR YOU,

1401
00:50:52,583 --> 00:50:53,584
AND IF YOU'RE NOT SURE

1402
00:50:53,584 --> 00:50:55,152
ABOUT THE CONFIDENCE METRICS

1403
00:50:55,152 --> 00:50:57,254
AND SO FORTH, THE,

1404
00:50:57,254 --> 00:50:58,923
THE EUROPEAN BIOINFORMATICS

1405
00:50:58,923 --> 00:51:00,691
INSTITUTE HAS SOME VERY NICE

1406
00:51:00,691 --> 00:51:02,326
RESOURCES, INCLUDING,

1407
00:51:02,326 --> 00:51:04,361
A REALLY GREAT TUTORIAL

1408
00:51:04,361 --> 00:51:07,364
THAT THAT REALLY, GIVES,

1409
00:51:07,564 --> 00:51:09,366
A VERY NICE INTRODUCTION TO,

1410
00:51:09,366 --> 00:51:10,234
TO ALPHAFOLD

1411
00:51:10,234 --> 00:51:13,137
AND THE CONFIDENCE METRICS AND SO FORTH.

1412
00:51:13,137 --> 00:51:16,140
SO I'D LIKE TO END BY

1413
00:51:16,440 --> 00:51:19,209
REALLY JUST POINTING OUT THAT I THINK

1414
00:51:19,209 --> 00:51:21,111
WITH THE STRUCTURE PREDICTION REVOLUTION,

1415
00:51:21,111 --> 00:51:23,013
THERE IS A REALLY DRAMATIC

1416
00:51:23,013 --> 00:51:25,883
DEMOCRATIZATION OF STRUCTURAL BIOLOGY

1417
00:51:25,883 --> 00:51:27,618
THAT IS UNDERWAY.

1418
00:51:27,618 --> 00:51:29,420
I THINK UNTIL RECENTLY,

1419
00:51:29,420 --> 00:51:30,721
THERE WAS SORT OF THIS,

1420
00:51:30,721 --> 00:51:33,157
YOU KNOW, THIS FIREWALL BETWEEN,

1421
00:51:33,157 --> 00:51:34,892
THE PEOPLE WHO,

1422
00:51:34,892 --> 00:51:36,794
YOU KNOW, DO GENETICS AND CELL BIOLOGY

1423
00:51:36,794 --> 00:51:37,795
AND SOME BIOCHEMISTRY

1424
00:51:37,795 --> 00:51:41,432
AND THOSE WHO DO STRUCTURAL BIOLOGY AND

1425
00:51:41,432 --> 00:51:44,702
I THINK THAT WITH THE,

1426
00:51:45,202 --> 00:51:45,536
YOU KNOW,

1427
00:51:45,536 --> 00:51:46,804
THE ADVENT OF ALPHAFOLD

1428
00:51:46,804 --> 00:51:47,871
AND OTHER RELATED

1429
00:51:47,871 --> 00:51:49,573
ALGORITHMS, ALL OF US,

1430
00:51:49,573 --> 00:51:50,541
WHETHER WE'RE

1431
00:51:50,541 --> 00:51:54,478
GENETICISTS, CELL BIOLOGISTS, BIOCHEMISTS

1432
00:51:54,478 --> 00:51:55,813
OR SOMETHING ELSE,

1433
00:51:55,813 --> 00:51:57,781
WE CAN ALL NOW BEGIN

1434
00:51:57,781 --> 00:51:58,782
TO REALLY THINK

1435
00:51:58,782 --> 00:52:01,785
STRUCTURALLY ABOUT ALL OF THE PROBLEMS

1436
00:52:01,952 --> 00:52:03,420
THAT WE WORK ON.

1437
00:52:03,420 --> 00:52:04,555
AND,

1438
00:52:04,555 --> 00:52:05,222
SO I THINK,

1439
00:52:05,222 --> 00:52:06,323
I THINK THIS, THIS WALL

1440
00:52:06,323 --> 00:52:08,726
IS COMPLETELY BREAKING DOWN

1441
00:52:08,726 --> 00:52:11,495
AND, YOU KNOW, FOR MY LAB

1442
00:52:11,495 --> 00:52:12,830
AND I THINK FOR A LOT OF OTHER LABS

1443
00:52:12,830 --> 00:52:14,598
NOW, A REALLY TYPICAL

1444
00:52:14,598 --> 00:52:17,601
SORT OF WORKFLOW IS TO EITHER,

1445
00:52:17,868 --> 00:52:19,703
YOU KNOW, USE THE RESOURCES

1446
00:52:19,703 --> 00:52:21,805
THAT I MENTIONED A FEW SLIDES AGO,

1447
00:52:21,805 --> 00:52:25,175
OR ONE'S OWN STRUCTURE, PREDICTIONS TO,

1448
00:52:25,542 --> 00:52:28,612
TO IDENTIFY POTENTIAL NEW INTERACTIONS,

1449
00:52:28,946 --> 00:52:30,314
WHICH IN MANY CASES

1450
00:52:30,314 --> 00:52:32,416
LEAD TO A COMPELLING HYPOTHESIS

1451
00:52:32,416 --> 00:52:34,952
THAT IS IMMEDIATELY TESTABLE,

1452
00:52:34,952 --> 00:52:35,686
YOU KNOW, THROUGH SITE

1453
00:52:35,686 --> 00:52:37,521
DIRECTED MUTAGENESIS,

1454
00:52:37,521 --> 00:52:39,923
AND STRUCTURE FUNCTION ANALYSIS

1455
00:52:39,923 --> 00:52:40,624
WITH, OF COURSE,

1456
00:52:40,624 --> 00:52:43,627
THE ULTIMATE GOAL ALWAYS BEING TO CONFIRM

1457
00:52:43,627 --> 00:52:44,862
THESE PREDICTIONS

1458
00:52:44,862 --> 00:52:46,697
USING STRUCTURE DETERMINATION.

1459
00:52:46,697 --> 00:52:48,432
BUT WHAT'S WHAT'S SO POWERFUL

1460
00:52:48,432 --> 00:52:49,166
AND SO BEAUTIFUL

1461
00:52:49,166 --> 00:52:50,000
ABOUT THE STRUCTURE

1462
00:52:50,000 --> 00:52:51,969
PREDICTION, REVOLUTION

1463
00:52:51,969 --> 00:52:56,273
IS THAT WE CAN ALL, THINK STRUCTURALLY

1464
00:52:56,273 --> 00:53:00,110
AND WE CAN LONG BEFORE WE HAVE

1465
00:53:00,244 --> 00:53:01,578
THE REAGENTS

1466
00:53:01,578 --> 00:53:02,413
OR THE RESULTS

1467
00:53:02,413 --> 00:53:03,180
OF AN EXPERIMENTAL

1468
00:53:03,180 --> 00:53:04,481
STRUCTURE DETERMINATION,

1469
00:53:04,481 --> 00:53:07,384
WE CAN ACTUALLY, START,

1470
00:53:07,384 --> 00:53:09,686
FORMULATING AND TESTING HYPOTHESES.

1471
00:53:09,686 --> 00:53:11,155
AND, AND I ALSO WANT TO SAY

1472
00:53:11,155 --> 00:53:12,089
THAT, YOU KNOW, FAR

1473
00:53:12,089 --> 00:53:13,357
FROM SOMEHOW

1474
00:53:13,357 --> 00:53:15,092
PUTTING STRUCTURAL BIOLOGISTS

1475
00:53:15,092 --> 00:53:15,926
OUT OF BUSINESS,

1476
00:53:15,926 --> 00:53:16,894
I THINK IT'S IN,

1477
00:53:16,894 --> 00:53:17,561
IN MANY WAYS

1478
00:53:17,561 --> 00:53:19,897
HAVING THE OPPOSITE EFFECT BECAUSE,

1479
00:53:19,897 --> 00:53:21,999
SO MANY MORE GROUPS CAN NOW,

1480
00:53:22,966 --> 00:53:24,968
FORMULATE

1481
00:53:24,968 --> 00:53:27,404
STRUCTURAL HYPOTHESES, YOU KNOW,

1482
00:53:27,404 --> 00:53:31,875
FEASIBLE STRUCTURAL HYPOTHESES THAT THAT,

1483
00:53:32,309 --> 00:53:32,976
WE'RE NOW

1484
00:53:32,976 --> 00:53:34,845
TALKING TO THE STRUCTURAL BIOLOGISTS

1485
00:53:34,845 --> 00:53:37,681
ABOUT AND GETTING THEIR HELP TESTING.

1486
00:53:37,681 --> 00:53:39,183
SO, SO FOR ME,

1487
00:53:39,183 --> 00:53:39,450
YOU KNOW,

1488
00:53:39,450 --> 00:53:41,485
IT IS REALLY DRAMATICALLY STRENGTHENED

1489
00:53:41,485 --> 00:53:43,020
MY INTERACTION

1490
00:53:43,020 --> 00:53:45,089
WITH WITH LOCAL STRUCTURE

1491
00:53:45,089 --> 00:53:47,491
DETERMINATION GROUPS.

1492
00:53:47,491 --> 00:53:48,492
I DON'T WANT TO LEAVE YOU

1493
00:53:48,492 --> 00:53:50,260
WITH TWO ROSY OF A PICTURE.

1494
00:53:50,260 --> 00:53:53,564
I WILL POINT OUT THAT THAT ALPHAFOLD

1495
00:53:53,564 --> 00:53:54,765
IS NOT INFALLIBLE.

1496
00:53:54,765 --> 00:53:55,866
THERE ARE LOTS OF FALSE

1497
00:53:55,866 --> 00:53:57,367
NEGATIVE INTERACTIONS

1498
00:53:57,367 --> 00:53:59,403
AND STILL SOME FALSE POSITIVES.

1499
00:53:59,403 --> 00:54:02,606
THERE ARE ERRORS IN IN RESIDUE

1500
00:54:02,606 --> 00:54:03,774
AND DOMAIN POSITIONING

1501
00:54:03,774 --> 00:54:06,810
AND OF COURSE, A LIMITED SORT OF,

1502
00:54:07,044 --> 00:54:10,347
SURVEY OF THE CONFORMATIONAL DYNAMICS.

1503
00:54:10,347 --> 00:54:12,249
SO ALL OF THIS MEANS THAT,

1504
00:54:12,249 --> 00:54:13,150
AND TOGETHER WITH,

1505
00:54:13,150 --> 00:54:13,984
WITH WHAT I SAID

1506
00:54:13,984 --> 00:54:14,852
JUST A MOMENT AGO,

1507
00:54:14,852 --> 00:54:15,853
ALL OF THIS MEANS THAT

1508
00:54:15,853 --> 00:54:17,454
THAT EXPERIMENTAL STRUCTURE

1509
00:54:17,454 --> 00:54:19,256
DETERMINATION IS STILL REALLY

1510
00:54:19,256 --> 00:54:21,658
AN ABSOLUTELY ESSENTIAL,

1511
00:54:21,658 --> 00:54:25,229
PART OF OF DERIVING,

1512
00:54:25,395 --> 00:54:26,864
A DEEP MECHANISTIC

1513
00:54:26,864 --> 00:54:28,332
UNDERSTANDING OF ALL THE PROCESSES

1514
00:54:28,332 --> 00:54:29,333
THAT WE CARE ABOUT.

1515
00:54:29,333 --> 00:54:30,400
SO I'M GOING TO STOP THERE

1516
00:54:30,400 --> 00:54:33,103
AND THANK THE PEOPLE WHO DID THE WORK.

1517
00:54:33,103 --> 00:54:34,404
SO TYCHO,

1518
00:54:34,404 --> 00:54:36,240
REALLY FABULOUS POSTDOC

1519
00:54:36,240 --> 00:54:37,875
WHO IS ON THE JOB MARKET.

1520
00:54:37,875 --> 00:54:43,580
DID ALL THE EXPERIMENTAL, WORK ON TC-NER

1521
00:54:43,580 --> 00:54:46,583
ERNST REALLY LED THE WAY ON THE,

1522
00:54:46,817 --> 00:54:48,218
STRUCTURE PREDICTION

1523
00:54:48,218 --> 00:54:50,154
SIDE WITH HELP FROM FROM HELEN

1524
00:54:50,154 --> 00:54:51,955
AND JACK, TWO UNDERGRADUATES.

1525
00:54:51,955 --> 00:54:53,857
AND I WANT TO THANK MY AMAZING

1526
00:54:53,857 --> 00:54:56,093
COLLABORATOR LUCAS AND HIS TRAINEES

1527
00:54:57,194 --> 00:54:58,896
FOR, JUST BLOWING

1528
00:54:58,896 --> 00:55:00,931
OUR MINDS WITH WITH HOW QUICKLY

1529
00:55:00,931 --> 00:55:03,867
THEY CAN, ACTUALLY GENERATE

1530
00:55:03,867 --> 00:55:06,870
EXPERIMENTAL STRUCTURE.

1531
00:55:07,237 --> 00:55:10,274
DETERMINATION TO TO HELP US TEST

1532
00:55:10,440 --> 00:55:11,909
WHETHER WE'RE ON THE RIGHT TRACK

1533
00:55:11,909 --> 00:55:14,878
WITH OUR, WITH OUR PREDICTIONS.

1534
00:55:14,878 --> 00:55:17,281
AND I THINK I'M GOING TO STOP THERE AND,

1535
00:55:17,281 --> 00:55:18,916
JUST THANK MY FUNDING SOURCES,

1536
00:55:18,916 --> 00:55:19,349
OF COURSE,

1537
00:55:19,349 --> 00:55:21,251
AND GIVE ANOTHER SHOUT OUT TO NVIDIA

1538
00:55:21,251 --> 00:55:25,556
FOR GENEROUSLY, DONATING GPU TIME TO,

1539
00:55:26,490 --> 00:55:28,959
START, START ZIPPING THROUGH,

1540
00:55:28,959 --> 00:55:31,128
THESE MANY PREDICTIONS THAT WE WANT TO,

1541
00:55:31,128 --> 00:55:32,563
TO GENERATE TO,

1542
00:55:32,563 --> 00:55:35,566
TO WORK TOWARDS A STRUCTURAL INTERACTOME.

1543
00:55:36,967 --> 00:55:37,701
ALL RIGHT.

1544
00:55:37,701 --> 00:55:39,770
SO, I'LL STOP THERE.

1545
00:55:39,770 --> 00:55:42,773
I GUESS I'LL STOP SHARING.

1546
00:55:44,808 --> 00:55:45,909
THANKS SO MUCH, JOHANNES.

1547
00:55:45,909 --> 00:55:48,912
THAT WAS WONDERFUL.

1548
00:55:49,279 --> 00:55:49,846
JIM. THANK YOU.

1549
00:55:49,846 --> 00:55:51,748
CHUNZHANG DO YOU WANT TO
START WITH THE QUESTIONS.

1550
00:55:51,748 --> 00:55:52,482
YEAH, SURE.

1551
00:55:52,482 --> 00:55:54,284
SO I'LL GO THROUGH THE,

1552
00:55:54,284 --> 00:55:56,353
QUESTIONS IN THE CHAT BOX.

1553
00:55:56,353 --> 00:55:59,523
SO WE HAVE A QUESTION FROM, MARCELLA.

1554
00:55:59,590 --> 00:56:00,757
SAGAR.

1555
00:56:00,757 --> 00:56:01,725
WHAT IS THE

1556
00:56:01,725 --> 00:56:03,460
ADVANTAGE OF USING ALPHAFOLD

1557
00:56:03,460 --> 00:56:04,194
AND MULTIMODE,

1558
00:56:04,194 --> 00:56:06,563
TO IDENTIFY

1559
00:56:06,563 --> 00:56:08,298
PROTEIN-PROTEIN INTERACTIONS.

1560
00:56:08,298 --> 00:56:09,299
ALPHAFOLD MODEL,

1561
00:56:09,299 --> 00:56:12,102
POLYMER ASSUME THE PROTEIN IS UNFOLDED

1562
00:56:12,102 --> 00:56:13,637
WITHOUT ITS PROTEIN

1563
00:56:13,637 --> 00:56:14,471
BINDING PARTNER

1564
00:56:14,471 --> 00:56:16,840
AND FOLDED INVOLVES INTERACTION

1565
00:56:16,840 --> 00:56:18,408
WITH THE BAIT PROTEIN.

1566
00:56:18,408 --> 00:56:19,810
COULD YOU IDENTIFY A NOVEL

1567
00:56:19,810 --> 00:56:21,078
BINDING PARTNER ASSUMING

1568
00:56:21,078 --> 00:56:22,913
THE PROTEIN IS ALREADY FOLDING

1569
00:56:22,913 --> 00:56:24,748
AND THE OTHER FOLDING PROTEIN

1570
00:56:24,748 --> 00:56:25,782
WITH THE BAIT PROTEIN?

1571
00:56:28,218 --> 00:56:30,220
I'M NOT SURE I

1572
00:56:30,220 --> 00:56:33,457
ENTIRELY UNDERSTAND THE QUESTION.

1573
00:56:33,924 --> 00:56:36,860
I GUESS

1574
00:56:36,860 --> 00:56:40,430
I'LL SAY THAT IN MOST CASES, BOTH,

1575
00:56:41,064 --> 00:56:43,700
YOU KNOW, BOTH THE BOTH THE BAIT AND

1576
00:56:43,700 --> 00:56:47,871
AND THE PREY PROTEINS WILL FOLD,

1577
00:56:48,605 --> 00:56:50,774
YOU KNOW, BOTH THERMODYNAMICALLY

1578
00:56:50,774 --> 00:56:54,878
AS WELL AS WITH ALPHAFOLD ON THEIR OWN.

1579
00:56:56,446 --> 00:56:59,449
THERE ARE INSTANCES SO

1580
00:56:59,816 --> 00:57:02,052
SO THERE ARE INTRINSICALLY DISORDERED

1581
00:57:02,052 --> 00:57:03,920
REGIONS, AS EVERYONE KNOWS.

1582
00:57:03,920 --> 00:57:05,856
AND A LOT OF THESE CONTAIN

1583
00:57:05,856 --> 00:57:09,026
SHORT LINEAR MOTIFS, SHORT PEPTIDES

1584
00:57:09,192 --> 00:57:10,861
THAT ARE USUALLY HIGHLY CONSERVED

1585
00:57:10,861 --> 00:57:12,763
BUT EMBEDDED WITHIN OTHERWISE

1586
00:57:12,763 --> 00:57:14,197
DISORDERED REGIONS.

1587
00:57:14,197 --> 00:57:17,601
AND, WE OFTEN SEE

1588
00:57:17,701 --> 00:57:19,603
THAT THESE ARE UNFOLD

1589
00:57:19,603 --> 00:57:21,004
FOLDED ON THEIR OWN.

1590
00:57:21,004 --> 00:57:21,972
BUT WHEN WE FOLD THEM

1591
00:57:21,972 --> 00:57:23,473
WITH A FUNCTIONALLY RELEVANT PARTNER

1592
00:57:23,473 --> 00:57:27,010
THAT THEY NOW ADOPT, A HELICAL FOLD.

1593
00:57:27,644 --> 00:57:28,945
SO, SO THAT IS,

1594
00:57:28,945 --> 00:57:29,346
I THINK

1595
00:57:29,346 --> 00:57:30,814
PROBABLY THE MOST PROMINENT

1596
00:57:30,814 --> 00:57:34,518
SORT OF SITUATION IN WHICH THE, YOU KNOW,

1597
00:57:34,518 --> 00:57:35,319
THE, THE,

1598
00:57:35,319 --> 00:57:36,586
THE ACT OF ACTUALLY

1599
00:57:36,586 --> 00:57:39,289
BRINGING TWO PARTNERS TOGETHER

1600
00:57:39,289 --> 00:57:40,457
INDUCES FOLDING,

1601
00:57:40,457 --> 00:57:41,591
BUT OTHERWISE,

1602
00:57:41,591 --> 00:57:44,594
THAT THAT IS TYPICALLY NOT THE CASE.

1603
00:57:44,594 --> 00:57:46,663
BUT BUT I'M NOT SURE I ENTIRELY

1604
00:57:46,663 --> 00:57:48,098
UNDERSTOOD THE QUESTION, BUT I HOPE,

1605
00:57:48,098 --> 00:57:49,533
HOPEFULLY I GOT AT LEAST PART OF IT.

1606
00:57:53,203 --> 00:57:54,504
I'LL ASK THE NEXT ONE.

1607
00:57:54,504 --> 00:57:57,641
WEI YANG (HI WEI) ASKS

1608
00:57:58,842 --> 00:57:59,142
IS 

1609
00:57:59,142 --> 00:58:01,478
STK19 IS PREDICTED TO INTERACT WITH MSH2

1610
00:58:01,478 --> 00:58:04,481
TO MUCH TO JUST BELOW CSA AND XPD

1611
00:58:04,715 --> 00:58:05,582
WHAT IS KNOWN OF 

1612
00:58:05,582 --> 00:58:08,585
STK19 AND MISMATCH REPAIR?

1613
00:58:09,152 --> 00:58:11,655
NOTHING TO MY KNOWLEDGE.

1614
00:58:11,655 --> 00:58:13,824
AND,

1615
00:58:13,824 --> 00:58:16,927
I DON'T REMEMBER THE EXACT SPOC SCORE,

1616
00:58:16,927 --> 00:58:17,494
BUT I THINK

1617
00:58:17,494 --> 00:58:17,994
I THINK THERE WAS

1618
00:58:17,994 --> 00:58:20,397
A BIT OF A DROP OFF IN TERMS OF,

1619
00:58:20,397 --> 00:58:21,832
YOU KNOW, THE,

1620
00:58:21,832 --> 00:58:25,302
THE ACTUAL CONFIDENCE VALUE.

1621
00:58:27,371 --> 00:58:28,872
BUT, YOU KNOW,

1622
00:58:28,872 --> 00:58:29,706
IT'S POSSIBLE,

1623
00:58:29,706 --> 00:58:32,042
I DON'T THINK ANYBODY HAS LOOKED AT

1624
00:58:32,042 --> 00:58:35,112
WHETHER THERE IS A FUNCTIONAL INTERPLAY

1625
00:58:35,112 --> 00:58:38,815
OF STK19 WITH MISMATCH REPAIR.

1626
00:58:39,383 --> 00:58:42,386
AND THIS IS, YOU KNOW, THIS IS WHERE

1627
00:58:43,186 --> 00:58:46,423
THERE IS OBVIOUSLY OFTEN A GRAY ZONE,

1628
00:58:46,923 --> 00:58:48,158
WITH THIS STRUCTURES

1629
00:58:48,158 --> 00:58:49,926
PREDICTION BUSINESS, THE

1630
00:58:49,926 --> 00:58:51,661
I WOULD SAY THE MOST,

1631
00:58:51,661 --> 00:58:55,232
YOU KNOW, COMPELLING CASES ARE

1632
00:58:55,232 --> 00:58:57,901
WHEN YOU GET A PREDICTION

1633
00:58:57,901 --> 00:58:59,136
THAT REALLY GIVES YOU

1634
00:58:59,136 --> 00:59:00,604
A EUREKA MOMENT, RIGHT?

1635
00:59:00,604 --> 00:59:03,306
AS IN THE CASE OF STK19

1636
00:59:03,306 --> 00:59:06,676
INTERACTING WITH CSA AND XPD,

1637
00:59:07,077 --> 00:59:08,378
TWO FACTORS

1638
00:59:08,378 --> 00:59:10,514
ALREADY KNOWN TO BE IN THE SAME PATHWAY.

1639
00:59:10,514 --> 00:59:12,582
THAT'S THAT'S REALLY KIND OF A NO BRAINER

1640
00:59:12,582 --> 00:59:13,116
WHERE YOU

1641
00:59:13,116 --> 00:59:14,084
YOU LOOK AT THAT PREDICTION,

1642
00:59:14,084 --> 00:59:14,951
YOU SAY,

1643
00:59:14,951 --> 00:59:17,220
YOU KNOW THAT THAT IS

1644
00:59:17,220 --> 00:59:18,722
THAT MAKES SO MUCH SENSE

1645
00:59:18,722 --> 00:59:20,957
THAT I'M GOING TO SPEND

1646
00:59:20,957 --> 00:59:23,427
THE TIME IT TAKES TO TEST THIS IDEA.

1647
00:59:24,961 --> 00:59:27,230
IN OTHER INSTANCES, IT CAN BE

1648
00:59:27,230 --> 00:59:30,400
OBVIOUSLY A MUCH RISKIER PROPOSITION.

1649
00:59:30,834 --> 00:59:32,602
AND SO IN THE CASE OF STK

1650
00:59:32,602 --> 00:59:35,605
19 AND MISMATCH REPAIR, YOU KNOW, I CAN'T

1651
00:59:35,772 --> 00:59:37,441
I REALLY CAN'T COMMENT,

1652
00:59:37,441 --> 00:59:40,143
ON WHETHER THAT'S WORTH PURSUING OR NOT

1653
00:59:40,143 --> 00:59:41,478
OR WHETHER THERE'S ANY,

1654
00:59:41,478 --> 00:59:44,080
ANY PRIOR EVIDENCE.

1655
00:59:44,080 --> 00:59:47,083
WE CERTAINLY DON'T HAVE ANY.

1656
00:59:47,918 --> 00:59:48,251
OKAY.

1657
00:59:48,251 --> 00:59:51,221
THE NEXT QUESTION IS FROM DMITRI GORDENIN.

1658
00:59:51,655 --> 00:59:52,389
WOULD YOU,

1659
00:59:52,389 --> 00:59:53,890
WOULD YOUR PLASMID SYSTEM

1660
00:59:53,890 --> 00:59:55,525
BE APPLICABLE TO NON-BULKY

1661
00:59:55,525 --> 00:59:56,693
 BASED LESIONS

1662
00:59:56,693 --> 00:59:59,696
IMPEDING RNA POLYMERASE?

1663
01:00:01,097 --> 01:00:04,100
RIGHT.

1664
01:00:04,901 --> 01:00:07,871
I THINK I THINK IN PRINCIPLE, YES.

1665
01:00:08,071 --> 01:00:08,839
AND IT'S

1666
01:00:08,839 --> 01:00:12,342
ACTUALLY BRINGS UP AN INTERESTING POINT,

1667
01:00:13,276 --> 01:00:15,745
WHICH IS THAT THERE ARE PRESUMABLY

1668
01:00:15,745 --> 01:00:16,246
THERE ARE

1669
01:00:16,246 --> 01:00:16,847
AND I'M SURE

1670
01:00:16,847 --> 01:00:17,814
THERE ARE PROBABLY OTHER PEOPLE

1671
01:00:17,814 --> 01:00:18,281
ON THE CALL

1672
01:00:18,281 --> 01:00:19,916
WHO KNOW MORE ABOUT THIS THAN I DO,

1673
01:00:19,916 --> 01:00:23,019
BUT THERE ARE PROBABLY LESIONS, SUCH

1674
01:00:23,019 --> 01:00:26,022
AS ABASIC SITES THAT WOULDN'T TRIGGER

1675
01:00:26,256 --> 01:00:28,859
THE GLOBAL GENOMIC REPAIR PATHWAY,

1676
01:00:28,859 --> 01:00:31,728
BUT THAT WOULD ABSOLUTELY STALL

1677
01:00:31,728 --> 01:00:32,929
RNA POLYMERASE.

1678
01:00:34,130 --> 01:00:35,632
BUT THEN I GUESS BUT THEN I GUESS

1679
01:00:35,632 --> 01:00:36,266
THE ISSUE IS

1680
01:00:36,266 --> 01:00:39,269
THAT THEY MAY NOT STALL THE XPD HELICASE.

1681
01:00:39,436 --> 01:00:43,707
SO, YEAH, IT'S AN INTERESTING QUESTION.

1682
01:00:43,707 --> 01:00:45,675
WE'VE NOT DONE ANYTHING BEYOND

1683
01:00:45,675 --> 01:00:48,812
USING THE CISPLATIN 1,3-G13

1684
01:00:49,379 --> 01:00:51,481
INTRA STRAND CROSS-LINKED.

1685
01:00:51,481 --> 01:00:52,482
BUT CERTAINLY

1686
01:00:52,482 --> 01:00:55,585
THE SYSTEM LENDS ITSELF TO EXPLORING

1687
01:00:56,520 --> 01:01:00,724
MORE WIDELY WHAT RANGE OF LESIONS

1688
01:01:00,724 --> 01:01:03,760
WOULD ACTUALLY LEAD TO PRODUCTIVE,

1689
01:01:03,760 --> 01:01:06,763
TECHNIQUE FOR.

1690
01:01:07,631 --> 01:01:07,898
OKAY.

1691
01:01:07,898 --> 01:01:09,432
DMITRI HAS A FOLLOW UP QUESTION

1692
01:01:09,432 --> 01:01:11,501
ON A COMPLETELY DIFFERENT TOPIC

1693
01:01:11,501 --> 01:01:12,702
HOW MUCH ALPHAFOLD BASED

1694
01:01:12,702 --> 01:01:14,004
PIPELINES ARE APPLICABLE

1695
01:01:14,004 --> 01:01:16,273
TO PREDICTION OF SNP EFFECTS PER SE?

1696
01:01:16,273 --> 01:01:16,973
WELL INTEGRATED

1697
01:01:16,973 --> 01:01:18,308
WITH EXISTING BIOINFORMATICS

1698
01:01:18,308 --> 01:01:21,311
PREDICTION METHODS, POLYPHEN, ETC..

1699
01:01:21,912 --> 01:01:22,178
RIGHT?

1700
01:01:22,178 --> 01:01:22,512
I THINK WELL,

1701
01:01:22,512 --> 01:01:26,416
I THINK THE UNDERLYING QUESTION HERE IS,

1702
01:01:26,416 --> 01:01:29,419
WHETHER YOU CAN USE ALPHAFOLD TO,

1703
01:01:29,619 --> 01:01:33,056
YOU KNOW, MODEL POINT MUTATIONS.

1704
01:01:33,490 --> 01:01:37,460
AND BY AND LARGE, ALPHAFOLD REALLY

1705
01:01:38,495 --> 01:01:39,963
IGNORES IF YOU.

1706
01:01:39,963 --> 01:01:41,531
SO IF YOU INTRODUCE, YOU KNOW,

1707
01:01:42,499 --> 01:01:45,502
A FEW MUTATIONS,

1708
01:01:45,869 --> 01:01:48,872
IN YOUR, YOUR BAIT OR YOUR PREY PROTEIN,

1709
01:01:49,139 --> 01:01:50,974
ALPHAFOLD WILL LARGELY IGNORE THAT.

1710
01:01:50,974 --> 01:01:51,942
AND THE REASON FOR THAT IS,

1711
01:01:51,942 --> 01:01:53,009
OF COURSE, THAT

1712
01:01:53,009 --> 01:01:55,946
WHAT IT DOES IS IT IT CONSIDERS,

1713
01:01:55,946 --> 01:01:56,580
YOU KNOW,

1714
01:01:56,580 --> 01:01:59,583
A WHOLE MULTIPLE SEQUENCE ALIGNMENT.

1715
01:01:59,583 --> 01:02:01,484
SO IT IT ALWAYS LOOKS FOR,

1716
01:02:01,484 --> 01:02:03,153
YOU KNOW, AS MANY,

1717
01:02:03,153 --> 01:02:05,188
ORTHOLOGS AS IT CAN,

1718
01:02:05,188 --> 01:02:06,556
AND THEN IT BUILDS,

1719
01:02:06,556 --> 01:02:08,692
YOU KNOW, MULTIPLE SEQUENCE ALIGNMENT.

1720
01:02:08,692 --> 01:02:10,393
IT LOOKS FOR, FOR COVARIATION.

1721
01:02:10,393 --> 01:02:13,463
AND SO IF, IF, IF YOU TWEAK ONE

1722
01:02:14,230 --> 01:02:16,366
PARTICULAR MEMBER OF THE MSA,

1723
01:02:16,366 --> 01:02:18,335
IT REALLY HAS NO EFFECT.

1724
01:02:18,335 --> 01:02:19,903
BUT OF COURSE THERE ARE

1725
01:02:19,903 --> 01:02:21,838
THERE ARE ALGORITHMS

1726
01:02:21,838 --> 01:02:24,074
THAT HAVE BEEN DEVELOPED

1727
01:02:24,074 --> 01:02:26,810
TO, ADDRESS THAT VERY THING.

1728
01:02:26,810 --> 01:02:32,415
AND ALPHA ALPHA MISSENSE IS PROBABLY,

1729
01:02:32,849 --> 01:02:33,850
YOU KNOW, THE BEST,

1730
01:02:33,850 --> 01:02:37,087
I THINK STILL RESOURCE TO ACTUALLY,

1731
01:02:37,754 --> 01:02:40,924
GET A SENSE OF WHETHER A MUTATION IS, IS

1732
01:02:40,924 --> 01:02:42,258
DELETERIOUS OR NEUTRAL.

1733
01:02:45,128 --> 01:02:47,664
NOW THE NEXT QUESTION IS FROM ROB SCOTT.

1734
01:02:47,664 --> 01:02:48,832
HELLO.

1735
01:02:48,832 --> 01:02:50,634
THANKS VERY MUCH FOR THE TALK.

1736
01:02:50,634 --> 01:02:51,401
AND,

1737
01:02:51,401 --> 01:02:53,970
HE'S WORKING IN, ALL ATOM

1738
01:02:53,970 --> 01:02:56,373
MOLECULAR DYNAMICS SIMULATION,

1739
01:02:56,373 --> 01:02:58,141
AND HE'S LOOKING FOR APPLICATION

1740
01:02:58,141 --> 01:02:59,476
IN DNA REPAIR.

1741
01:02:59,476 --> 01:03:00,343
AND HE ASK,

1742
01:03:00,343 --> 01:03:01,845
CAN YOU PLEASE TELL

1743
01:03:01,845 --> 01:03:03,780
IS THERE SOME PARTICULAR PROBLEM

1744
01:03:03,780 --> 01:03:05,348
SEE, YOU COULD RECOMMEND THAT

1745
01:03:05,348 --> 01:03:08,351
THAT HE WORK ON.

1746
01:03:10,453 --> 01:03:11,154
AND FOR AN ALL

1747
01:03:11,154 --> 01:03:14,157
ATOM SIMULATION.

1748
01:03:15,258 --> 01:03:16,393
GOSH.

1749
01:03:16,393 --> 01:03:17,127
WELL,

1750
01:03:17,127 --> 01:03:19,062
I MEAN, I THINK,

1751
01:03:19,062 --> 01:03:21,865
I THINK AS SOON AS YOU START TO MODEL,

1752
01:03:21,865 --> 01:03:23,533
YOU KNOW, NOT ONLY PROTEINS, BUT,

1753
01:03:23,533 --> 01:03:27,137
BUT OTHER, YOU KNOW, LIGANDS,

1754
01:03:27,637 --> 01:03:30,140
YOU HAVE TO START TO MOVE INTO AN ALL

1755
01:03:30,140 --> 01:03:33,410
ATOM, SORT OF SETTING.

1756
01:03:33,410 --> 01:03:34,811
AND OF COURSE,

1757
01:03:34,811 --> 01:03:35,578
THAT'S A LOT OF

1758
01:03:35,578 --> 01:03:37,881
WHAT ALPHAFOLD THREE IS DOING.

1759
01:03:37,881 --> 01:03:39,716
BUT I DON'T

1760
01:03:39,716 --> 01:03:41,718
I DON'T THINK I HAVE A REALLY GOOD

1761
01:03:41,718 --> 01:03:44,721
SPECIFIC, RECOMMENDATION FOR YOU.

1762
01:03:45,121 --> 01:03:46,222
I'M HAPPY TO CORRESPOND

1763
01:03:46,222 --> 01:03:47,891
WITH YOU OVER THAT. AND IF YOU WANT.

1764
01:03:49,926 --> 01:03:50,627
THERE'S A

1765
01:03:50,627 --> 01:03:51,761
QUESTION FROM

1766
01:03:51,761 --> 01:03:52,929
SOMEBODY THAT JUST LABELED

1767
01:03:52,929 --> 01:03:53,797
WOODGATES LAB.

1768
01:03:53,797 --> 01:03:54,664
I DON'T KNOW IF IT'S ROGER

1769
01:03:54,664 --> 01:03:56,999
OR SOMEBODY ELSE SAYING NICE.

1770
01:03:56,999 --> 01:03:58,601
NICE TALK AND DETAIL.

1771
01:03:58,601 --> 01:03:59,803
SMILEY FACE.

1772
01:03:59,803 --> 01:04:00,970
DID YOU CHECK THE CURRENT

1773
01:04:00,970 --> 01:04:02,005
ACTIVITY OF STK

1774
01:04:02,005 --> 01:04:03,406
19 IN THE COMPLEX,

1775
01:04:03,406 --> 01:04:06,242
OR IS IT ACTING MORE AS AN ADAPTER?

1776
01:04:06,242 --> 01:04:09,612
WELL, SO SO I CERTAINLY HAVE NOT

1777
01:04:09,612 --> 01:04:12,115
WE'VE NOT LOOKED AT, KINASE ACTIVITY,

1778
01:04:12,115 --> 01:04:13,450
BUT REALLY, I MEAN,

1779
01:04:13,450 --> 01:04:14,117
IT DOESN'T LOOK

1780
01:04:14,117 --> 01:04:16,686
LIKE ANYTHING LIKE A KINASE.

1781
01:04:16,686 --> 01:04:18,188
SO IT WOULD BE TRULY SHOCKING

1782
01:04:18,188 --> 01:04:19,456
IF IT HAD KINASE ACTIVITY.

1783
01:04:19,456 --> 01:04:20,924
BUT I THINK, YES,

1784
01:04:20,924 --> 01:04:23,893
ACTUALLY, PUBLISHED A PAPER

1785
01:04:24,127 --> 01:04:25,695
WHERE HE ADDRESSED

1786
01:04:25,695 --> 01:04:26,496
NOT ONLY THE MIS-

1787
01:04:26,496 --> 01:04:27,797
ANNOTATION OF STK19,

1788
01:04:27,797 --> 01:04:28,965
BUT ALSO,

1789
01:04:28,965 --> 01:04:30,233
YOU KNOW,

1790
01:04:30,233 --> 01:04:31,968
THE FACT THAT IT'S NOT A KINASE.

1791
01:04:31,968 --> 01:04:33,436
AND HE DID NOT LOOK AT IT

1792
01:04:33,436 --> 01:04:35,839
IN THE CONTEXT OF A COMPLEX,

1793
01:04:35,839 --> 01:04:37,540
BUT CERTAINLY IN ISOLATION,

1794
01:04:37,540 --> 01:04:39,876
AND FOUND NO EVIDENCE OF KINASE ACTIVITY.

1795
01:04:39,876 --> 01:04:41,344
AND IT WOULD BE VERY SHOCKING

1796
01:04:41,344 --> 01:04:44,347
IF IT SUDDENLY BECAME A KINASE,

1797
01:04:44,414 --> 01:04:47,150
WHEN IT WHEN IT BINDS TO THE TC-NER COMPLEX

1798
01:04:49,085 --> 01:04:52,422
AND WE HAVE A MESSAGE FROM, FROM,

1799
01:04:52,422 --> 01:04:53,690
ANDREW JACKSON

1800
01:04:53,690 --> 01:04:56,593
AND HE SAY, HI, YOHANNES IN SPOC

1801
01:04:56,593 --> 01:04:57,727
DID YOU LOOK AT,

1802
01:04:57,727 --> 01:04:59,162
WHICH PARAMETER GAVE MOST

1803
01:04:59,162 --> 01:05:00,463
PREDICTIVE VALUE?

1804
01:05:00,463 --> 01:05:03,433
IF SO, WHICH WERE MOST INFORMATIVE?

1805
01:05:03,600 --> 01:05:05,735
RIGHT. YEAH, THAT'S A GREAT QUESTION.

1806
01:05:05,735 --> 01:05:08,972
SO I'M TRYING TO REMEMBER

1807
01:05:08,972 --> 01:05:11,975
FROM THE STRUCTURAL FEATURES,

1808
01:05:12,709 --> 01:05:16,479
IT WAS ACTUALLY,

1809
01:05:16,479 --> 01:05:18,214
I BELIEVE IT WAS A METRIC

1810
01:05:18,214 --> 01:05:20,750
THAT WE CALL AVERAGE MODELS.

1811
01:05:20,750 --> 01:05:23,386
SO, SO ONE OF THE HOMEGROWN METRICS

1812
01:05:23,386 --> 01:05:24,420
THAT WE CAME UP WITH

1813
01:05:24,420 --> 01:05:27,957
WAS SIMPLY TO QUANTIFY THE AGREEMENT

1814
01:05:28,057 --> 01:05:30,760
AMONG THE FIVE INDEPENDENTLY

1815
01:05:30,760 --> 01:05:32,562
TRAINED ALPHAFOLD MODELS.

1816
01:05:32,562 --> 01:05:34,597
SO PROBABLY MANY OF YOU KNOW

1817
01:05:34,597 --> 01:05:37,600
THAT THAT ALPHAFOLD WAS TRAINED ON FIVE,

1818
01:05:38,568 --> 01:05:41,571
SLIGHTLY DIFFERENT, REGIMES.

1819
01:05:41,571 --> 01:05:43,706
AND SO WHEN YOU GIVE IT A PREDICTION

1820
01:05:43,706 --> 01:05:45,408
WHERE YOU GIVE IT A PROTEIN

1821
01:05:45,408 --> 01:05:46,242
OR A PROTEIN PAIR,

1822
01:05:46,242 --> 01:05:48,511
IT ACTUALLY SPITS OUT FIVE SOLUTIONS.

1823
01:05:48,511 --> 01:05:50,747
AND AND SO WHAT WE FOUND

1824
01:05:50,747 --> 01:05:52,282
WAS THAT THE GREATER

1825
01:05:52,282 --> 01:05:52,949
THE AGREEMENT

1826
01:05:52,949 --> 01:05:53,917
AMONG THOSE,

1827
01:05:53,917 --> 01:05:57,287
FIVE MODELS, THE MORE LIKELY

1828
01:05:57,854 --> 01:05:59,155
SOMETHING IS TO BE TRUE.

1829
01:05:59,155 --> 01:05:59,989
AND WE ACTUALLY HAVE

1830
01:05:59,989 --> 01:06:01,591
THAT AS A FEATURE IN SPOC.

1831
01:06:01,591 --> 01:06:03,059
SO, SO ON THE STRUCTURAL SIDE,

1832
01:06:03,059 --> 01:06:05,895
THAT IS ACTUALLY ONE OF THE BEST FEATURES

1833
01:06:05,895 --> 01:06:08,998
ON THE BIOLOGICAL SIDE, IT TURNS OUT

1834
01:06:08,998 --> 01:06:12,001
ACTUALLY TO BE BIOGRID.

1835
01:06:12,235 --> 01:06:14,804
SO AS YOU KNOW, BIOGRID,

1836
01:06:14,804 --> 01:06:16,406
YOU KNOW, IT CURATES

1837
01:06:16,406 --> 01:06:19,542
NOT ONLY THE LITERATURE BUT ALSO,

1838
01:06:19,809 --> 01:06:22,745
YOU KNOW, A LOT OF PROTEOMIC DATABASES

1839
01:06:22,745 --> 01:06:26,115
AND SO, PERHAPS NOT SURPRISINGLY, BIO

1840
01:06:26,115 --> 01:06:29,686
GRID IS IN FACT, THE MOST,

1841
01:06:30,453 --> 01:06:34,257
IMPORTANT METRIC ON THE BIOLOGICAL SIDE.

1842
01:06:34,958 --> 01:06:37,961
BUT IT IS BY NO MEANS IT IS BY NO MEANS,

1843
01:06:38,394 --> 01:06:41,397
YOU KNOW, SUFFICIENT.

1844
01:06:41,664 --> 01:06:44,601
AND, AND AND WE ACTUALLY SEE

1845
01:06:44,601 --> 01:06:47,704
MANY EXAMPLES OF PROTEIN PAIRS

1846
01:06:47,704 --> 01:06:49,606
LIKE STK19-XPD,

1847
01:06:49,606 --> 01:06:51,474
WHICH ARE NOT REPORTED IN BIOGRID.

1848
01:06:51,474 --> 01:06:54,043
THAT INTERACTION IS COMPLETELY UNKNOWN.

1849
01:06:54,043 --> 01:06:56,713
YET THEY SCORE SUPER HIGHLY.

1850
01:06:56,713 --> 01:06:59,682
SO SO, YOU KNOW, IT'S A COMPLEX

1851
01:06:59,682 --> 01:07:00,917
IT'S A COMPLEX PICTURE.

1852
01:07:01,918 --> 01:07:04,420
SO, YOU KNOW, THE CRISPR SCREENS AND,

1853
01:07:04,420 --> 01:07:05,088
YOU KNOW,

1854
01:07:05,088 --> 01:07:06,489
CO-EXPRESSION

1855
01:07:06,489 --> 01:07:08,291
ALL OF THOSE THINGS CONTRIBUTE.

1856
01:07:08,291 --> 01:07:08,825
AND SO

1857
01:07:08,825 --> 01:07:09,559
AND SO IT'S

1858
01:07:09,559 --> 01:07:11,160
REALLY A SUM OF MANY PARTS

1859
01:07:11,160 --> 01:07:12,161
THAT GIVES YOU THE,

1860
01:07:12,161 --> 01:07:13,463
THE ULTIMATE PERFORMANCE.

1861
01:07:13,463 --> 01:07:16,032
THAT IS SPOC.

1862
01:07:16,032 --> 01:07:16,599
OKAY.

1863
01:07:16,599 --> 01:07:18,167
ROB SCOTT JUST SAYS THANKS.

1864
01:07:18,167 --> 01:07:18,835
HE WAS GUY

1865
01:07:18,835 --> 01:07:19,402
WITH THE QUESTION

1866
01:07:19,402 --> 01:07:20,703
ABOUT WHAT PROBLEMS TO WORK ON.

1867
01:07:20,703 --> 01:07:22,005
MAYBE THIS IS SOMETHING YOU COULD FOLLOW

1868
01:07:22,005 --> 01:07:25,508
UP, OFFLINE, BUT KEN WANTS TO ANSWER

1869
01:07:25,508 --> 01:07:27,310
ASK A QUESTION KEN.

1870
01:07:27,310 --> 01:07:28,978
WHAT A WONDERFUL PRESENTATION.

1871
01:07:28,978 --> 01:07:30,546
I REALLY ENJOYED THAT.

1872
01:07:30,546 --> 01:07:33,549
IF SDK19 IS INVOLVED IN

1873
01:07:33,716 --> 01:07:35,418
DNA REPAIR AND TRANSCRIPTION

1874
01:07:35,418 --> 01:07:36,519
COUPLED REPAIR.

1875
01:07:36,519 --> 01:07:38,388
AND WE KNOW THAT THE

1876
01:07:38,388 --> 01:07:41,491
MUTATIONS IN THAT, CAUSE HUMAN DISEASES

1877
01:07:42,258 --> 01:07:45,428
ARE, THE TRANSCRIPTION COUPLED REPAIR

1878
01:07:46,362 --> 01:07:47,897
ABNORMALITIES SEEM TO INVOLVE

1879
01:07:47,897 --> 01:07:49,766
MORE NEUROLOGIC DEGENERATION

1880
01:07:49,766 --> 01:07:51,401
RATHER THAN CANCER.

1881
01:07:51,401 --> 01:07:53,136
I HAVE YOU BEEN ABLE

1882
01:07:53,136 --> 01:07:55,471
TO LOOK AT ANY OF THESE NEW LARGE

1883
01:07:55,471 --> 01:07:57,140
GENOME DATABASES

1884
01:07:57,140 --> 01:07:58,274
TO SEE IF THERE ARE

1885
01:07:58,274 --> 01:08:00,176
ANY ABNORMALITIES IN STK

1886
01:08:00,176 --> 01:08:03,379
19 THAT READ HUMAN DISEASE OR TO CATCH.

1887
01:08:04,047 --> 01:08:06,182
YEAH. THAT'S A IT'S A TERRIFIC QUESTION.

1888
01:08:06,182 --> 01:08:08,351
PERSONALLY WE'VE NOT DONE THAT.

1889
01:08:08,351 --> 01:08:10,420
I BET,

1890
01:08:10,420 --> 01:08:11,921
I BET SOME OF THE OTHER GROUPS

1891
01:08:11,921 --> 01:08:15,024
WORKING ON STK19 THAT ARE MORE

1892
01:08:15,024 --> 01:08:17,126
GENETICALLY ORIENTED THAN WE ARE,

1893
01:08:17,126 --> 01:08:18,661
HAVE PROBABLY DONE THAT.

1894
01:08:18,661 --> 01:08:19,929
AND I IMAGINE

1895
01:08:19,929 --> 01:08:22,498
THAT THE ABSENCE OF, OF ANY,

1896
01:08:22,498 --> 01:08:26,202
STATEMENTS OR REPORTS IN, IN ANY OF THE,

1897
01:08:26,602 --> 01:08:30,640
PAPERS,

1898
01:08:30,640 --> 01:08:32,742
SUGGESTS THAT SO FAR AT LEAST,

1899
01:08:32,742 --> 01:08:34,744
THERE ARE NO PATIENT MUTATIONS.

1900
01:08:34,744 --> 01:08:36,245
AND I NOW REALIZE THAT I,

1901
01:08:36,245 --> 01:08:37,914
I THINK I MAY HAVE ACTUALLY,

1902
01:08:37,914 --> 01:08:41,217
NEGLECTED TO MENTION THE OTHER GROUPS.

1903
01:08:41,684 --> 01:08:42,618
I THINK I HAD THEIR,

1904
01:08:42,618 --> 01:08:44,654
THEIR REFERENCES ON MY SLIDE,

1905
01:08:44,654 --> 01:08:46,389
BUT I DO WANT TO POINT OUT THAT,

1906
01:08:46,389 --> 01:08:50,026
I THINK IT'S

1907
01:08:50,326 --> 01:08:53,296
THREE OTHER GROUPS ALSO REPORTED

1908
01:08:53,296 --> 01:08:56,899
THE ROLE OF STK19 IN TC-NER

1909
01:08:57,133 --> 01:08:58,935
AND, AND A NUMBER OF THOSE,

1910
01:08:58,935 --> 01:09:00,870
ALSO PROVIDED EVIDENCE

1911
01:09:00,870 --> 01:09:03,006
FOR THE INTERACTION WITH XPD,

1912
01:09:04,273 --> 01:09:05,441
IN ALL CASES BASED ON

1913
01:09:05,441 --> 01:09:08,411
THE ALPHAFOLD PREDICTION.

1914
01:09:11,414 --> 01:09:12,982
I DON'T SEE ADDITIONAL QUESTIONS.

1915
01:09:12,982 --> 01:09:16,686
SO I THINK WE'LL JUST, STOP HERE

1916
01:09:16,686 --> 01:09:18,755
AND I WOULD LIKE TO TAKE THIS OPPORTUNITY

1917
01:09:18,755 --> 01:09:20,590
TO THANK DOCTOR WALTER

1918
01:09:20,590 --> 01:09:21,858
FOR SUCH A FASCINATING

1919
01:09:21,858 --> 01:09:23,693
AND INSIGHTFUL PRESENTATION

1920
01:09:23,693 --> 01:09:24,894
AND THE GROUNDBREAKING WORK THAT

1921
01:09:24,894 --> 01:09:25,928
CLEARLY ADVANCED

1922
01:09:25,928 --> 01:09:27,697
UNDERSTANDING FOR TRANSCRIPTION

1923
01:09:27,697 --> 01:09:29,399
COUPLED WITH DNA REPAIR.

1924
01:09:29,399 --> 01:09:31,134
I WOULD ALSO LIKE TO THANK EVERYONE

1925
01:09:31,134 --> 01:09:31,634
FOR JOINING

1926
01:09:31,634 --> 01:09:33,936
TODAY'S SESSION AND THANK YOU ALL.

1927
01:09:33,936 --> 01:09:35,071
AND WE ARE LOOKING FORWARD

1928
01:09:35,071 --> 01:09:36,873
TO SEE YOU IN THE FUTURE EVENTS.

1929
01:09:36,873 --> 01:09:38,441
THANK YOU. HAVE A GREAT DAY.

1930
01:09:38,441 --> 01:09:39,542
THANK YOU, DOCTOR WALTER.

1931
01:09:39,542 --> 01:09:42,545
YEAH, THANK YOU ALL FOR LISTENING.

1932
01:09:42,912 --> 01:09:43,646
YEAH. BYE BYE.