1
00:00:00,850 --> 00:00:02,753
So, has everybody heard of CRISPR?

2
00:00:03,883 --> 00:00:06,485
I would be shocked if you hadn't.

3
00:00:06,509 --> 00:00:09,711
This is a technology --
it's for genome editing --

4
00:00:09,735 --> 00:00:12,569
and it's so versatile and so controversial

5
00:00:12,593 --> 00:00:15,834
that it's sparking all sorts
of really interesting conversations.

6
00:00:16,631 --> 00:00:18,694
Should we bring back the woolly mammoth?

7
00:00:19,281 --> 00:00:21,575
Should we edit a human embryo?

8
00:00:22,041 --> 00:00:24,163
And my personal favorite:

9
00:00:24,997 --> 00:00:28,648
How can we justify
wiping out an entire species

10
00:00:28,672 --> 00:00:31,116
that we consider harmful to humans

11
00:00:31,140 --> 00:00:32,442
off the face of the Earth,

12
00:00:32,466 --> 00:00:33,983
using this technology?

13
00:00:35,165 --> 00:00:38,448
This type of science
is moving much faster

14
00:00:38,472 --> 00:00:41,316
than the regulatory mechanisms
that govern it.

15
00:00:41,340 --> 00:00:43,407
And so, for the past six years,

16
00:00:43,431 --> 00:00:45,376
I've made it my personal mission

17
00:00:45,863 --> 00:00:48,873
to make sure that as many people
as possible understand

18
00:00:48,897 --> 00:00:51,997
these types of technologies
and their implications.

19
00:00:52,021 --> 00:00:56,550
Now, CRISPR has been the subject
of a huge media hype,

20
00:00:57,089 --> 00:01:01,448
and the words that are used most often
are "easy" and "cheap."

21
00:01:02,337 --> 00:01:05,476
So what I want to do is drill down
a little bit deeper

22
00:01:05,500 --> 00:01:09,684
and look into some of the myths
and the realities around CRISPR.

23
00:01:10,954 --> 00:01:12,904
If you're trying to CRISPR a genome,

24
00:01:13,594 --> 00:01:16,194
the first thing that you have to do
is damage the DNA.

25
00:01:17,098 --> 00:01:20,140
The damage comes in the form
of a double-strand break

26
00:01:20,164 --> 00:01:21,728
through the double helix.

27
00:01:21,752 --> 00:01:24,504
And then the cellular repair
processes kick in,

28
00:01:25,088 --> 00:01:27,727
and then we convince
those repair processes

29
00:01:27,751 --> 00:01:29,627
to make the edit that we want,

30
00:01:30,054 --> 00:01:31,596
and not a natural edit.

31
00:01:31,620 --> 00:01:32,865
That's how it works.

32
00:01:33,929 --> 00:01:35,652
It's a two-part system.

33
00:01:35,676 --> 00:01:39,047
You've got a Cas9 protein
and something called a guide RNA.

34
00:01:39,071 --> 00:01:41,533
I like to think of it as a guided missile.

35
00:01:41,557 --> 00:01:44,203
So the Cas9 --
I love to anthropomorphize --

36
00:01:44,227 --> 00:01:47,457
so the Cas9 is kind of this Pac-Man thing

37
00:01:47,481 --> 00:01:49,026
that wants to chew DNA,

38
00:01:49,050 --> 00:01:53,206
and the guide RNA is the leash
that's keeping it out of the genome

39
00:01:53,230 --> 00:01:56,079
until it finds the exact spot
where it matches.

40
00:01:56,912 --> 00:01:59,806
And the combination of those two
is called CRISPR.

41
00:01:59,830 --> 00:02:01,398
It's a system that we stole

42
00:02:01,422 --> 00:02:04,316
from an ancient, ancient
bacterial immune system.

43
00:02:05,469 --> 00:02:09,209
The part that's amazing about it
is that the guide RNA,

44
00:02:10,041 --> 00:02:11,932
only 20 letters of it,

45
00:02:11,956 --> 00:02:13,594
are what target the system.

46
00:02:14,570 --> 00:02:16,713
This is really easy to design,

47
00:02:16,737 --> 00:02:18,556
and it's really cheap to buy.

48
00:02:18,985 --> 00:02:22,990
So that's the part
that is modular in the system;

49
00:02:23,014 --> 00:02:24,812
everything else stays the same.

50
00:02:25,481 --> 00:02:28,912
This makes it a remarkably easy
and powerful system to use.

51
00:02:30,047 --> 00:02:34,287
The guide RNA and the Cas9
protein complex together

52
00:02:34,311 --> 00:02:36,243
go bouncing along the genome,

53
00:02:36,267 --> 00:02:39,760
and when they find a spot
where the guide RNA matches,

54
00:02:39,784 --> 00:02:42,639
then it inserts between the two strands
of the double helix,

55
00:02:42,663 --> 00:02:44,231
it rips them apart,

56
00:02:44,692 --> 00:02:47,380
that triggers the Cas9 protein to cut,

57
00:02:47,962 --> 00:02:49,381
and all of a sudden,

58
00:02:49,816 --> 00:02:51,716
you've got a cell that's in total panic

59
00:02:51,740 --> 00:02:54,365
because now it's got a piece
of DNA that's broken.

60
00:02:55,000 --> 00:02:56,296
What does it do?

61
00:02:56,320 --> 00:02:58,514
It calls its first responders.

62
00:02:58,959 --> 00:03:01,581
There are two major repair pathways.

63
00:03:01,605 --> 00:03:06,674
The first just takes the DNA
and shoves the two pieces back together.

64
00:03:06,698 --> 00:03:08,796
This isn't a very efficient system,

65
00:03:08,820 --> 00:03:11,549
because what happens is
sometimes a base drops out

66
00:03:11,573 --> 00:03:13,000
or a base is added.

67
00:03:13,024 --> 00:03:16,817
It's an OK way to maybe, like,
knock out a gene,

68
00:03:16,841 --> 00:03:20,074
but it's not the way that we really want
to do genome editing.

69
00:03:20,098 --> 00:03:22,993
The second repair pathway
is a lot more interesting.

70
00:03:23,017 --> 00:03:24,653
In this repair pathway,

71
00:03:24,677 --> 00:03:27,358
it takes a homologous piece of DNA.

72
00:03:27,382 --> 00:03:30,022
And now mind you, in a diploid
organism like people,

73
00:03:30,046 --> 00:03:34,294
we've got one copy of our genome
from our mom and one from our dad,

74
00:03:34,318 --> 00:03:35,593
so if one gets damaged,

75
00:03:35,617 --> 00:03:38,012
it can use the other
chromosome to repair it.

76
00:03:38,036 --> 00:03:39,669
So that's where this comes from.

77
00:03:40,518 --> 00:03:41,982
The repair is made,

78
00:03:42,006 --> 00:03:43,957
and now the genome is safe again.

79
00:03:44,616 --> 00:03:46,139
The way that we can hijack this

80
00:03:46,497 --> 00:03:50,205
is we can feed it a false piece of DNA,

81
00:03:50,229 --> 00:03:52,373
a piece that has homology on both ends

82
00:03:52,397 --> 00:03:54,096
but is different in the middle.

83
00:03:54,120 --> 00:03:56,587
So now, you can put
whatever you want in the center

84
00:03:56,611 --> 00:03:58,126
and the cell gets fooled.

85
00:03:58,150 --> 00:04:00,269
So you can change a letter,

86
00:04:00,293 --> 00:04:01,558
you can take letters out,

87
00:04:01,582 --> 00:04:04,506
but most importantly,
you can stuff new DNA in,

88
00:04:04,530 --> 00:04:06,279
kind of like a Trojan horse.

89
00:04:07,089 --> 00:04:09,274
CRISPR is going to be amazing,

90
00:04:09,298 --> 00:04:12,916
in terms of the number of different
scientific advances

91
00:04:12,940 --> 00:04:14,597
that it's going to catalyze.

92
00:04:14,621 --> 00:04:17,842
The thing that's special about it
is this modular targeting system.

93
00:04:17,866 --> 00:04:21,633
I mean, we've been shoving DNA
into organisms for years, right?

94
00:04:21,657 --> 00:04:23,784
But because of the modular
targeting system,

95
00:04:23,808 --> 00:04:26,233
we can actually put it
exactly where we want it.

96
00:04:27,423 --> 00:04:33,092
The thing is that there's
a lot of talk about it being cheap

97
00:04:33,116 --> 00:04:34,858
and it being easy.

98
00:04:34,882 --> 00:04:37,694
And I run a community lab.

99
00:04:38,242 --> 00:04:41,798
I'm starting to get emails from people
that say stuff like,

100
00:04:41,822 --> 00:04:44,219
"Hey, can I come to your open night

101
00:04:44,243 --> 00:04:47,860
and, like, maybe use CRISPR
and engineer my genome?"

102
00:04:47,884 --> 00:04:48,994
(Laugher)

103
00:04:49,018 --> 00:04:50,519
Like, seriously.

104
00:04:51,376 --> 00:04:53,179
I'm, "No, you can't."

105
00:04:53,203 --> 00:04:54,213
(Laughter)

106
00:04:54,237 --> 00:04:56,592
"But I've heard it's cheap.
I've heard it's easy."

107
00:04:56,616 --> 00:04:58,799
We're going to explore that a little bit.

108
00:04:58,823 --> 00:05:00,772
So, how cheap is it?

109
00:05:00,796 --> 00:05:03,206
Yeah, it is cheap in comparison.

110
00:05:03,665 --> 00:05:07,284
It's going to take the cost of the average
materials for an experiment

111
00:05:07,308 --> 00:05:09,840
from thousands of dollars
to hundreds of dollars,

112
00:05:09,864 --> 00:05:11,800
and it cuts the time a lot, too.

113
00:05:11,824 --> 00:05:13,904
It can cut it from weeks to days.

114
00:05:14,246 --> 00:05:15,738
That's great.

115
00:05:15,762 --> 00:05:18,452
You still need a professional lab
to do the work in;

116
00:05:18,476 --> 00:05:21,988
you're not going to do anything meaningful
outside of a professional lab.

117
00:05:22,012 --> 00:05:24,016
I mean, don't listen to anyone who says

118
00:05:24,040 --> 00:05:26,771
you can do this sort of stuff
on your kitchen table.

119
00:05:27,421 --> 00:05:31,929
It's really not easy
to do this kind of work.

120
00:05:31,953 --> 00:05:34,261
Not to mention,
there's a patent battle going on,

121
00:05:34,285 --> 00:05:36,111
so even if you do invent something,

122
00:05:36,135 --> 00:05:42,771
the Broad Institute and UC Berkeley
are in this incredible patent battle.

123
00:05:42,795 --> 00:05:45,183
It's really fascinating
to watch it happen,

124
00:05:45,207 --> 00:05:48,455
because they're accusing each other
of fraudulent claims

125
00:05:48,479 --> 00:05:50,210
and then they've got people saying,

126
00:05:50,234 --> 00:05:53,018
"Oh, well, I signed
my notebook here or there."

127
00:05:53,042 --> 00:05:55,143
This isn't going to be settled for years.

128
00:05:55,167 --> 00:05:56,327
And when it is,

129
00:05:56,351 --> 00:05:59,636
you can bet you're going to pay someone
a really hefty licensing fee

130
00:05:59,660 --> 00:06:01,009
in order to use this stuff.

131
00:06:01,343 --> 00:06:03,124
So, is it really cheap?

132
00:06:03,148 --> 00:06:08,347
Well, it's cheap if you're doing
basic research and you've got a lab.

133
00:06:09,220 --> 00:06:11,496
How about easy?
Let's look at that claim.

134
00:06:12,417 --> 00:06:14,905
The devil is always in the details.

135
00:06:15,881 --> 00:06:19,012
We don't really know
that much about cells.

136
00:06:19,036 --> 00:06:20,706
They're still kind of black boxes.

137
00:06:20,730 --> 00:06:25,590
For example, we don't know
why some guide RNAs work really well

138
00:06:25,614 --> 00:06:27,677
and some guide RNAs don't.

139
00:06:27,701 --> 00:06:31,169
We don't know why some cells
want to do one repair pathway

140
00:06:31,193 --> 00:06:33,651
and some cells would rather do the other.

141
00:06:34,270 --> 00:06:35,554
And besides that,

142
00:06:35,578 --> 00:06:38,447
there's the whole problem
of getting the system into the cell

143
00:06:38,471 --> 00:06:39,735
in the first place.

144
00:06:39,759 --> 00:06:41,761
In a petri dish, that's not that hard,

145
00:06:41,785 --> 00:06:44,230
but if you're trying to do it
on a whole organism,

146
00:06:44,254 --> 00:06:45,797
it gets really tricky.

147
00:06:46,224 --> 00:06:49,410
It's OK if you use something
like blood or bone marrow --

148
00:06:49,434 --> 00:06:51,661
those are the targets
of a lot of research now.

149
00:06:51,685 --> 00:06:53,936
There was a great story
of some little girl

150
00:06:53,960 --> 00:06:55,612
who they saved from leukemia

151
00:06:55,636 --> 00:06:58,310
by taking the blood out, editing it,
and putting it back

152
00:06:58,334 --> 00:07:00,383
with a precursor of CRISPR.

153
00:07:00,869 --> 00:07:03,665
And this is a line of research
that people are going to do.

154
00:07:03,689 --> 00:07:06,242
But right now, if you want to get
into the whole body,

155
00:07:06,266 --> 00:07:08,410
you're probably going
to have to use a virus.

156
00:07:08,434 --> 00:07:10,807
So you take the virus,
you put the CRISPR into it,

157
00:07:10,831 --> 00:07:12,537
you let the virus infect the cell.

158
00:07:12,561 --> 00:07:14,704
But now you've got this virus in there,

159
00:07:14,728 --> 00:07:17,416
and we don't know what the long-term
effects of that are.

160
00:07:17,440 --> 00:07:19,708
Plus, CRISPR has some off-target effects,

161
00:07:19,732 --> 00:07:22,693
a very small percentage,
but they're still there.

162
00:07:22,717 --> 00:07:25,515
What's going to happen
over time with that?

163
00:07:26,039 --> 00:07:28,251
These are not trivial questions,

164
00:07:28,275 --> 00:07:30,915
and there are scientists
that are trying to solve them,

165
00:07:30,939 --> 00:07:33,225
and they will eventually,
hopefully, be solved.

166
00:07:33,249 --> 00:07:36,883
But it ain't plug-and-play,
not by a long shot.

167
00:07:36,907 --> 00:07:38,666
So: Is it really easy?

168
00:07:39,032 --> 00:07:43,365
Well, if you spend a few years
working it out in your particular system,

169
00:07:43,389 --> 00:07:44,814
yes, it is.

170
00:07:45,426 --> 00:07:47,510
Now the other thing is,

171
00:07:47,534 --> 00:07:53,893
we don't really know that much about how
to make a particular thing happen

172
00:07:53,917 --> 00:07:56,822
by changing particular spots
in the genome.

173
00:07:57,306 --> 00:07:59,473
We're a long way away from figuring out

174
00:07:59,497 --> 00:08:01,866
how to give a pig wings, for example.

175
00:08:02,364 --> 00:08:05,228
Or even an extra leg -- I'd settle
for an extra leg.

176
00:08:05,252 --> 00:08:06,898
That would be kind of cool, right?

177
00:08:06,922 --> 00:08:08,458
But what is happening

178
00:08:08,482 --> 00:08:12,832
is that CRISPR is being used
by thousands and thousands of scientists

179
00:08:12,856 --> 00:08:15,228
to do really, really important work,

180
00:08:15,252 --> 00:08:20,696
like making better models
of diseases in animals, for example,

181
00:08:20,720 --> 00:08:25,702
or for taking pathways
that produce valuable chemicals

182
00:08:25,726 --> 00:08:29,608
and getting them into industrial
production in fermentation vats,

183
00:08:30,021 --> 00:08:33,482
or even doing really basic research
on what genes do.

184
00:08:34,022 --> 00:08:36,951
This is the story of CRISPR
we should be telling,

185
00:08:36,975 --> 00:08:40,439
and I don't like it
that the flashier aspects of it

186
00:08:40,463 --> 00:08:42,219
are drowning all of this out.

187
00:08:42,243 --> 00:08:46,817
Lots of scientists did a lot of work
to make CRISPR happen,

188
00:08:46,841 --> 00:08:48,460
and what's interesting to me

189
00:08:48,484 --> 00:08:52,994
is that these scientists
are being supported by our society.

190
00:08:53,423 --> 00:08:54,582
Think about it.

191
00:08:54,606 --> 00:08:58,625
We've got an infrastructure that allows
a certain percentage of people

192
00:08:58,983 --> 00:09:02,292
to spend all their time doing research.

193
00:09:02,984 --> 00:09:06,355
That makes us all the inventors of CRISPR,

194
00:09:06,998 --> 00:09:11,466
and I would say that makes us all
the shepherds of CRISPR.

195
00:09:11,490 --> 00:09:13,297
We all have a responsibility.

196
00:09:13,749 --> 00:09:17,705
So I would urge you to really learn
about these types of technologies,

197
00:09:18,010 --> 00:09:20,029
because, really, only in that way

198
00:09:20,415 --> 00:09:24,767
are we going to be able to guide
the development of these technologies,

199
00:09:24,791 --> 00:09:26,723
the use of these technologies

200
00:09:26,747 --> 00:09:30,502
and make sure that, in the end,
it's a positive outcome --

201
00:09:31,034 --> 00:09:34,166
for both the planet and for us.

202
00:09:34,698 --> 00:09:35,890
Thanks.

203
00:09:35,914 --> 00:09:39,816
(Applause)