00:00hi and welcome to the a 16z podcast I'm
00:03Hannah today's episode features a
00:05special conversation with renowned
00:07scientist George Church known for his
00:09groundbreaking work and methods used for
00:11the first genome sequence and for his
00:13work in genome editing writing and
00:14recoating church's innovations have
00:17become an essential building block for
00:18most of the DNA sequencing methods and
00:20companies we see today he is joined in
00:22this conversation with a 16z bio general
00:25partner Jorge Conde who among other
00:27things founded a company with church out
00:29of the church lab the to take us on a
00:31wild journey into the scientists mind
00:33and work starting with what the leading
00:34pioneer in the space makes of where we
00:36are today with CRISPR especially given
00:38recent news about CRISPR babies in China
00:40then moving on to the broader
00:42implications of all that on a cultural
00:43level to finally what it takes to go
00:46from science fiction to lab to reality
00:48so let's start at the beginning if we
00:51were to bet ten years ago whether we'd
00:54a mammoth baby or a Neanderthal baby
00:57which would you have bet would have come
01:00though unquestionably a Christopher baby
01:02yeah I mean I was not I used technical
01:05leap they all involved societal and
01:09ethical questions but that one probably
01:13had the clearest path that no because
01:15there's such divergence of opinion
01:18somebody was going to do it and would
01:21you have expected that it would have
01:22been essentially a rogue effort versus a
01:25solo effort as it seems to have been the
01:27case in the China CRISPR I mean the
01:30earth rises solo ever I've seen the
01:31author list is quite long and I also
01:33find it unlikely that a government as
01:36technically astute and as engaged in
01:41observation would be unaware of such an
01:45important thing if I were technically
01:47astute government they're very limited
01:50number topics I would be paying
01:51attention to and these would be things
01:52like you know nuclear biological
01:56encryption and CRISPR it's a short list
02:00so I don't think it's so low so let's
02:02talk a little bit about the way it's
02:03been sort of positioned at least
02:05publicly can you describe a little bit
02:07what the experiment actually was what
02:12did the scientist or scientists do in
02:14particular case for the crisper baby
02:15I've actually seen a lot of the data and
02:18the free prints and this was a simple in
02:22a certain sense application of CRISPR to
02:28alleviate a potential for HIV infection
02:31you know 900,000 people die every year
02:34of HIV and this was an approach to it
02:38and they did it by knocking out the gene
02:40that encodes the HIV receptor on the
02:45it's a ccr5 ccr5 which has already been
02:49approved for FDA clinical trials for
02:52Sangamo and for editing in adults that
02:56have AIDS that's a different scenario
02:58but it vets many of the issues that come
03:01up as to whether this is a reasonable
03:03editing strategy so first of all people
03:06have described it as knocking out the
03:08gene other people have described it as
03:10editing the ccr5 gene I haven't seen the
03:12data what exactly was done to ccr5 right
03:16so what CRISPR does well it's often
03:18described as editing it really is
03:21damaging that's what it does it's not
03:23really that good at precision editing
03:25hopefully we'll there will be a good way
03:28in the future and so what it does is it
03:30knocks out genes and in this case that's
03:32exactly what you want is you want to
03:33knock out the ccr5 gene and there's
03:35precedent for it about up to 10% of
03:38certain parts of Europe have a double
03:40null to double know in this case basis
03:42obviously your function will cease you
03:44need really both non functional in order
03:46to be resistant to the virus and it
03:49doesn't make you resistance all viruses
03:50doesn't even make resistant to all HIV
03:52viruses but that's not the point is it's
03:56like a vaccine it makes resistant
03:57whatever you vaccinated to yes and
03:59analogies were made in the consenting
04:01for this between this and vaccination
04:04there is no good vaccine there is no
04:06cure for hiv/aids and right now if you
04:11get it and there there's a million
04:13people who have been affected if you get
04:16it you're doomed to a lifetime of
04:17combined antiretroviral therapy which is
04:21not the thing that you would wish to
04:22have if you had any choice sure while
04:25vaccines if there did exist one
04:28would be quite a good choice I mean so
04:30this is as close as you can get to a
04:32vaccine so I read that the double nulls
04:34for ccr5 have increased predisposition
04:38to West Nile that is correct so there is
04:40a risk for almost every preventive and
04:42therapy and this is the risk this case
04:44in most populations that's considered a
04:46smaller medical risk it's obviously a
04:49case by case for populations and
04:51individuals and there are undoubtedly
04:53other advantages and disadvantages and I
04:56may be taking you a little bit out of
04:57context here but I've heard you describe
04:59CRISPR as genetic vandalism you know so
05:03do you think that that's a good
05:04application for germline editing let's
05:07handle ISM in the sense that it can add
05:10or delete a small number of base pairs
05:12typically in the range of one to
05:15hundreds it's not going to do something
05:18really wacky except it may be some
05:21incredibly low frequency again no drug
05:24is without its side effects that's why
05:26there's all the fine print that
05:27accompanies all the blue drugs so I
05:30think in this case it is what you want
05:32is exactly what you want you want to
05:33destroy the ccr5 gene without destroying
05:36any adjacent genes and that's every
05:39allele that I've seen in the literature
05:40for ccr5 for there done in adults or
05:43done in tissue culture is what you would
05:46want so I've read in this case in the
05:48Chinese CRISPR baby publication that
05:50there is some mosaicism that he might
05:52not have functionally knocked out all of
05:55the ccr5 so is there any worry that
05:57after the post experiment that this
05:59particular child might still be at risk
06:02for HIV infection so first of all in the
06:04approved clinical trials on adults that
06:08there is a lot of mosaicism it's
06:12considered part of the clinical trial
06:13and maybe as little as 20% are properly
06:18edited meaning double nulls that's
06:21enough though because all the rest are
06:23wiped out by the virus and then the ones
06:25that are edited dominate the t-cell
06:28population so it's one way of thinking
06:30about it this is basically selection for
06:32the edited t-cell so you don't develop
06:34immunosuppression right so as long as
06:35there's a fair number of properly edited
06:38ones now on the other hand looking at
06:39the data I don't see that much evidence
06:41for mosaicism it's quite possible that
06:45what you see in the pre-implantation
06:48embryo when you select a few cells out
06:53is not representative the final and the
06:56final is less mosaic or maybe even non
06:59those okay so when they talk about a
07:01baby having mosaicism in the case of the
07:04CRISPR baby essentially what they're
07:06referring to is that there are some
07:07cells that will have edits and some
07:10cells that won't and so essentially that
07:12child may grow up to be a mosaic of two
07:16different or multiple different cell
07:18types correct and the same thing I
07:20should note is true for adult gene
07:21therapies is that whether it's done ex
07:24vivo or in vivo it uses results in a
07:26high level of mosaicism because the
07:28delivery is inefficient and it may even
07:31be the case that the germline has lower
07:33mosaicism we need more data great so the
07:36amount of off target and mosaicism so
07:39far for these two babies seems to be low
07:42but time will tell you know it could be
07:44that we're just lucky the same way that
07:48the first in vitro fertilization Louise
07:50Brown turned out just fine
07:52and so that greatly influenced it
07:54shouldn't've I mean it's an N of one
07:56we shouldn't have all said Oh IVF is
07:58perfect because we have one perfect baby
08:00- you're referencing test-tube hysteria
08:02around the 1578 which subsided it grew
08:07too much and it subsided too quickly
08:10based on end of one and I think here we
08:13have an N up to maybe a minute and
08:15there's gonna be a lot of attention paid
08:17to the actual outcomes rather how we got
08:20there hopefully if you had been in
08:22charge of the project would you have
08:24done ccr5 or is there another different
08:26obvious application that I must vary
08:29preclinical you know where as I create
08:31technologies it's been used by companies
08:33that I found they do the clinical trial
08:35so I would probably would not be doing a
08:36clinical trial at all those just to put
08:38it in context but in terms of choice of
08:41target I have said publicly already that
08:44targets that have been championed by the
08:46critics to the extent they champion
08:49anything or the ones that they present
08:50as possibilities or as higher priority
08:53although with great reservations
08:55even for those are things that are
08:58typical Mendelian diseases that is to
09:00say diseases are very severe and are
09:05predictably heritable which are things
09:09like hemoglobinopathies salla Semien
09:13suzi fibrosis and so forth ignoring the
09:16fact that if you're in an ivf PGD clinic
09:19anyway you can select to do your CRISPR
09:22editing of your Mendelian disease you
09:25could just do selection for most of
09:27these things so I think that's kind of
09:29like they're rationalizing their choice
09:32which you know in the same sense that
09:34they might feel is rationalizing to pick
09:36a more prominent disease but also I
09:38think and all the examples you just
09:39cited you would actually need to edit
09:41the gene to create function as opposed
09:43to knocking out as it was the case for
09:45ccr5 right and does something some of
09:47the critics might think that it's
09:48attractive that Christopher's
09:50inappropriate at this moment because it
09:51gives us because it's more time to think
09:53about it but in any case yeah I think
09:55that we want an example of a disease
09:58that is very common and most the gene
10:01therapies are rare whether they're
10:02editing or not and we want something
10:06and certainly HIV falls in that category
10:09so it struck me as a plausibly
10:12justifiable choice possibly more
10:15justifiable something that you can avoid
10:17with genetic counseling or with PGD IVF
10:21so IVF PGD stands for in vitro
10:24fertilization with prenatal genetic
10:27diagnosis so the diagnosis can
10:29essentially be done before you implant
10:30the embryo from an in-vitro
10:33into the mother and so by some people's
10:36definition that's still kind of a lab
10:38resource rather than a baby and those
10:40are typically used for Mendelian
10:43diseases meaning that you can see in the
10:46parents for example both parents could
10:48be unaffected carriers you can predict
10:51that 25% of their children will or their
10:54embryos in in vitro fertilization could
10:56be affected with a very serious disease
10:59so now that the genies out of the bottle
11:01we have the first CRISPR babies born
11:03first of all what was the role of ethics
11:08the first project in the ccr5 Chinese
11:10crisper baby project and what do you see
11:12is the role of ethicists going forward
11:15well so the National Academy of Sciences
11:17in the US and with participation from
11:19China and other countries in February
11:222017 came out with a report where they
11:23listed 10 items that would be
11:26recommendations for prerequisites for
11:30doing to remind editing in children I
11:32mean obviously you can do germline
11:34editing in animals or you can do it in
11:37cells in culture or even embryos in
11:40culture but actually implanting it and
11:41having children and a lot of these had
11:44ethical components many of them were
11:45very similar to what you would expect
11:47the FDA or the C FDA or the EMA to be
11:51these are all regulatory agencies around
11:54the world would recommend for any
11:56therapeutic clinical trial they should
11:58all be focused on safety and efficacy
12:01and ethics and that's what these 10
12:04items look like for germline as well
12:07do you suspect or do you expect I should
12:09say that we're gonna see more and more
12:11of these experiments going forward or do
12:13you think that after this first one
12:15going back to the IVF example do you
12:18think there will be a pause well there
12:20probably will be something that looks
12:22like a pause but it will probably be an
12:24acceleration so that's the same thing
12:27happened with a prominent DNA that was
12:28supposedly a moratorium but during that
12:31time I mean I was a first-hand observer
12:33my research went faster because people
12:36were building incredible facilities for
12:38containment and they had just
12:40state-of-the-art equipment that helped
12:41everything go faster in my opinion and I
12:43think the same thing is gone with almost
12:45every major ethical debate is it
12:48attracts attention attracts money
12:49whatever it is ethical at the time is
12:53accelerated and then so whenever we
12:55become comfortable with it all that
12:57acceleration clicks into place and it's
13:00as if there's been a steady growth that
13:03doesn't mean we should be in cautious on
13:05the contrary I'm very much Pro
13:08regulation I think that regulation is
13:10what saves us from solidified and Vioxx
13:13and hormone replacement therapy and so
13:16forth long term so I think we need to
13:19support our regulatory agencies around
13:21world they are not agents of slowing
13:24things down there I phase and so
13:25exposing things out yeah and I think
13:27what's pretty clear we're seeing that
13:28today in the regulatory environment
13:29certainly here in the US I mean we've
13:31got the first cell therapies the first
13:33gene therapies the first digital
13:34therapies it's a pretty remarkable
13:36moment from a regulatory standpoint
13:38first and I think they like new
13:40technologies more than like the old ones
13:42the old ones tend to fail
13:45because they aren't so incremental
13:46they're no longer compared to the
13:48placebo they're compared to whatever
13:49they're an increment over or whatever
13:51therapy already works and they often
13:53fail but health brand new category
13:55you know monoclonal antibodies or cell
13:58therapies or gene therapies those just
14:00like blow past and create all sorts of
14:03new improvements dramatic improvements
14:06in safety and efficacy so the FDA is
14:07that's their mandate is to cure people
14:10not to stop people from practicing
14:12medicine so if just to take that vein if
14:15we look forward what do you see as sort
14:17of the next non-incremental sort of step
14:20function change in the way we treat
14:22disease or manage disease or even
14:24diagnose disease well first of all if we
14:26started diagnosing that would be a
14:27really big thing I think it's really
14:30we're a population even worldwide they
14:33were under diagnosed there's a lot of
14:35very cost-effective diagnosis that
14:37partly because they're cost effective
14:39they're undervalued and the care
14:41providers are not compensated as much as
14:44some less effective but expensive
14:47medicine so that's one thing diagnosis
14:49would be terrific and that's part of
14:51preventive medicine so we talked a lot
14:52about precision medicine but the
14:54preventative part gets kind of swept
14:57under the rug a bit if you look at the
14:58pie charts for a number of government
15:01agencies including the NCI NIH general
15:05is preventive as sort of in the 1 to 5
15:07percent of the pie chart but it's
15:09payback is enormous and so basically
15:12you're saying misaligned incentives in
15:14human behavior has sort of mitigated how
15:16much prevention we actually do right but
15:18that would be a huge breakthroughs if we
15:20could do more diagnosis and more
15:22trenchant now the ultimate diagnosis for
15:24genetics is whole genome sequencing and
15:26environmental monitoring with sequencing
15:28as well for pathogens allergens and so
15:30forth the therapeutic cognate of that
15:34you know preventing serious Mendelian
15:37diseases that are very predictive and
15:40very often single gene or have enough of
15:42a single gene component that they're
15:44ready for medical practice thousands of
15:47them and those can be prevented we often
15:50talk about gene therapy actually that's
15:51a million-dollar drug it is once and
15:54done so you don't have a lifetime of
15:55those thing but it is expense that we
15:58need to acknowledge that partly because
16:00a lot of them are rare if you get a
16:02common gene therapy like let's say aging
16:05reversal or some major infectious agent
16:08that everybody wants to be vaccinated
16:09against it's like most infectious ages
16:12have potentially billions of customers
16:14and that bring the price down radically
16:16but in addition to gene therapy either
16:20in adults children fetuses or germline
16:22there is the option of doing IVF PGD
16:26that we already mentioned and even
16:28earlier and matchmaking so if you never
16:30meet or fall in love with someone who is
16:35predisposed to create heavily disease
16:38genetically diseased children
16:40that's very both cost effective and
16:43humane so you're describing 23andme
16:45meets 1000 I am not actually I'm
16:48describing a whole genome sequencing
16:50okay which is not there are a very small
16:53number of companies that provide whole
16:54genome sequencing because everything
16:56else anything less the whole genome
16:58sequencing is not medically powerful
17:00enough anything less than that missus
17:02gives you false assurance that combined
17:06with it some sort of dating that is an
17:08odd combination and possibly further
17:10combined with whoever's paying for the
17:12Mendelian cost right now which are about
17:15a million dollars per person doesn't
17:17have to be gene therapy which happens to
17:19be a million it can be just caregiving
17:21it adds up and somebody's paying for
17:24that typically insurers and employment
17:27benefits and they could be saving this
17:29money if they could encourage their
17:32clients patients to avoid following a
17:36lot marrying and having children when
17:38they have incompatible areas of someday
17:40and this actually works so Doria Shireen
17:42has eliminated significant and the only
17:46disease like tay-sachs
17:47by practicing a version of this probably
17:50isn't perfectly generalizable but there
17:52are verses this that that could keep a
17:54great deal of privacy and allow people
17:57to just never know whether they're
18:00affected or not whether there are
18:01carriers or not never know if anybody
18:03else is but still avoid meeting I mean
18:06the analog version of this was back in
18:08the day in certain communities Jewish
18:10communities where there was disease the
18:11rabbi would a sense the thorium was this
18:14exactly it was started by an individual
18:17who had five children row they were
18:19affected by the tay-sachs which is a
18:21terrible burden on the child and the
18:24family typically die before they're four
18:26years old they're painful and so he
18:29correctly determined that you could do
18:32this very inexpensively and mainly via
18:35matchmaking via matchmaking right so
18:37let's take another blast back to the
18:39past so about ten years ago you and I
18:41started a company in whole genome
18:44sequencing called gnome thank you called
18:46gnome not know me called gnome we used
18:50to have this constant back and forth but
18:52you thought it should be called no me I
18:53would call it no me I thought it should
18:54be called gnome and this was the market
18:56test more people listened to you than to
18:58me she's incredibly frustrating yeah and
19:01now I listen to you thank you this said
19:03no okay my rejoin her on that always was
19:06if you want to call it no me and I want
19:09to call you Jorge glasses and you were
19:11never a big fan of that one I don't got
19:13no problem with out they I think it's a
19:15better name it's nice thing that's gonna
19:17increase the brand it's just as more
19:19syllables as long but it just it rolls
19:22off the tongue goes yes so 10 years ago
19:24we basically made the bet that whole
19:27genome sequencing was important that
19:29interpretation of that data would be
19:30relevant that it would be meaningful 10
19:33years hence there still are not many
19:35people walking around that have had
19:37their whole genomes sequenced right
19:39despite the fact that the cost has now
19:40fallen arguably below $1,000 or at least
19:43we're at that thousand dollar threshold
19:45so I'd two questions for you number one
19:47is is the thousand dollar threshold for
19:50this to be useful for everyone to get
19:53sequenced too high a dollar number in
19:54other words does it need to be a hundred
19:56dollars or ten dollars mhm and number
19:58two to the extent that this hasn't
20:01happened yet why hasn't happened yet if
20:03it's not cost I would say there's three
20:06reasons why it hasn't happened yeah and
20:07I've been living this reality for most
20:10of my career I'm convinced that it would
20:12be valuable for the world it cost
20:14effective medicine preventive and I
20:17think the three reasons are one is cost
20:19the cost should probably be zero dollars
20:20secondly it's privacy we should have a
20:23convincing mechanism of people getting
20:26benefit from their genome without
20:28necessarily knowing their genome or
20:29anybody else knowing there's UNAM you
20:30can have something where it's only an
20:32encrypted form and not available to
20:34anybody including insurance and
20:36government a second and the third is
20:39most people don't understand the value
20:42it's either misrepresented but by both
20:44extremes so some people say oh it's so
20:47valuable that you're gonna whip out your
20:49cell phone and look at your genome twice
20:50a day and at the other extreme they say
20:52I can't imagine ever using it and the
20:55reality is somewhere in between and I
20:56think the analogy is seatbelts so seat
20:59belts were essentially free they were
21:01standard equipment they were required by
21:04law that you buckle and there were a lot
21:06of ad campaigns to get you to do so kind
21:08of like smoking and none of those were
21:11effective because people did you know
21:14the kind of ordinary math which is hey
21:16I've got a less than 1% chance of ever
21:18using a seat belt ever needing one so
21:21I'm not gonna bother and then the thing
21:23that made the difference was technology
21:24does since the buckling and turning off
21:27an annoying sound beam okay yeah okay so
21:29that's what made a difference and we
21:30need an equivalent thing it's a public
21:32health issue it's not an individual
21:34health issue so I don't benefit from
21:36being sequence they collect yeah most
21:38people 95 96 percent will get a blank
21:44sheet they should get a blank sheet in
21:46terms of really actionable very serious
21:49Mendelian diseases and that should be
21:51the expectation not the two extremes
21:53that you'll use it every day or that or
21:55everybody will use it every day or at
21:57the other extreme which is totally
21:58useless it's this strange thing where
22:01one to four percent of the population
22:03will have a very big impact on their
22:05life and the bottom line for their care
22:09providers millions of dollars huge
22:11impact on whole family
22:14if you're one of the unlucky 4% and we
22:18need to get that message out there and I
22:20think that bringing the price down to $0
22:22and and showing that it's protectable
22:24encrypted so that nobody can get access
22:27to it except for things that benefit you
22:29or your family or society that will get
22:32their attention but it's gonna take a
22:35little bit more than that's gonna take
22:36some anecdotes you would think that data
22:38would be better than anecdotes but you
22:39need both and I think it's gonna happen
22:42very soon now because we finally have
22:43the zero dollar genome and the
22:45encryption and we're starting to get
22:47communication of this rare advantage
22:50where you're not exempt even though the
22:53odds are that you're exempt
22:54you don't know that you're exempt until
22:56you get your genome sequence so two
22:58questions for you on the three ones
22:59you've laid out the first one is in the
23:02early days of Nome I remember when we
23:04would think about this question of
23:05security you correctly pointed out that
23:08if you really wanted my genome you would
23:11just wait for me to leave the room and
23:13collect it it's even more true than it
23:15was back in this chair off this table
23:18you know and you've got me got it so why
23:20is security and privacy is it even a
23:23meaningful thing to think about if it's
23:25an impossible thing to achieve well the
23:27point is if it's preventing people from
23:29getting their own genome sequence if
23:31they think that them seeing their own
23:33genome puts them at risk for somebody
23:36like hacking or requesting it or
23:38subpoenaing it and yes it's a problem
23:40because there is a difference between me
23:42willfully getting my genome and looking
23:44at it and somebody serously taking it
23:47okay so we can pass laws to punish
23:50people for surreptitiously taking our
23:51DNA we do have the genetic information
23:53non-discrimination Act of 2008 that is
23:56along those lines it's not perfect but
23:59it shows the intention of the public so
24:01that can kind of handle the abandoned
24:04DNA problem and we could keep shoring
24:06that up and building up those laws but
24:08then there's the question if I look at
24:10my genome if I have my genome available
24:12in text formats unprotected then anybody
24:15can come along and demand it right the
24:17insurance companies say I know you know
24:19it so I want to see it government cuz
24:21yeah I want to see it so I can convict
24:23your brother and if it's encrypted so
24:27Paquette then you can just say sorry
24:30it's out of my hands I don't have my
24:31genome if you want my genome you're
24:32gonna have to steal it from me right got
24:34it and I think that's where we are today
24:36finally by the way you may not remember
24:38this but we were laying out the risk
24:40factors and all of the other things for
24:41the consent for know on all the things
24:44that a potential recipient of their
24:47genome data would have to think about by
24:49far in a way my favorite one that you
24:51contributed was the potential risk that
24:53someone could plant your DNA at a crime
24:55scene right no high risk or low risk so
24:59that was also in a Personal Genome
25:00Project consent form which started
25:04around that same time as gnome did is it
25:06high risk or low risk you know I'd say
25:08that we're getting more more
25:09sophisticated at sequencing and
25:12methylation analysis you'd have to have
25:15the whole genome now rather than back
25:17then it might be just the CODIS parts
25:19that is the CODIS is just a few handful
25:22of simple sequence repeats that are used
25:25in criminal investigations like take
25:27forensics yeah forensic and you know CSI
25:30time stuff now you'd need the whole
25:33genome because if somebody felt it was
25:35being hacked they'd say well you know
25:37let's check the old you know a defense
25:38attorney could ask for the old you know
25:40further recast for methylation to show
25:42that it's the right age for example
25:43looking at my DNA from 20 years ago ah
25:45and you'd have to show or you could
25:48check the immune status so you could say
25:50oh you know does the immune status
25:52coincide with what which would be an
25:54argument for you to be like constantly
25:56sequencing your immune your blood DNA so
25:58you can date whatever samples you know
26:01yeah so for every hack there's a counter
26:04hack so I think I'm glad that we're not
26:06at that stage right at the moment even
26:08though we predicted it back in the you
26:10know 2005 so going back to 2005 can you
26:13describe briefly what the Personal
26:15Genome Project was because it was the
26:17first effort to really start to think
26:19through these issues the Personal Genome
26:21Project was one of the first
26:23recognitions about how identifiable both
26:26your genome is and also even parts of it
26:29and your medical records and people were
26:33starting to want to share genomic data
26:36ideally integrated so you could see what
26:41what we would now call precision
26:43medicine record would look like right
26:45back in 2005 and I wrote an editorial
26:48saying that this was a risk that the
26:51data could leak out and once it leaked
26:53out the people could be Rio Dental of
26:55their diseases could be determined from
26:57either the medical record or the genome
26:59or both and this has played out I mean
27:01there's many examples of millions of
27:04people being there medical records and
27:06or there's you know leaking out in
27:08various ways and of course now since
27:10then WikiLeaks has occurred which is
27:12just an example of how they can be
27:14officially stored publicly after leaking
27:17so I think that was what we were
27:19concerned about and we started the
27:21Personal Genome Project so he get people
27:23properly consented so they knew these
27:25risks they accepted them and you had to
27:27take a quiz right exactly up to that
27:29point many of the consent forms were
27:32long written in legalese a lot of
27:35language protected the institution
27:37rather than a person and you would sign
27:39them often under course of circumstances
27:42where you were afraid you weren't going
27:43to get the best medical care if you
27:46so we added to that a simple multi
27:49choice exam where you kind of
27:51simultaneously got educated if you
27:52didn't get a perfect score until you got
27:55a perfect score so it wasn't like we
27:57wanted 90% comprehension we want a 100%
27:59that you knew all of the risks and all
28:01the benefits and we had a record of that
28:03so that was where some of the key points
28:05the personal JUnit project was it but
28:07the other key point is we really wanted
28:08to share it not just what a lot of
28:10people call sharing even to this day you
28:12know 13 years later they call sharing
28:14medical data for research is really a
28:17silo that's hard to get into
28:20unfortunately it's not impossible to get
28:22into it's not really encrypted the way
28:23you would want it to be and so there's a
28:25lot of potential for leakage but it's
28:27hard enough for a regular scientist of
28:30good intention to get in to get access
28:32to it legitimately so we wanted
28:35something that's more like Wikipedia
28:36where you didn't have to agree to be a
28:38co-author on a paper you didn't have to
28:41pay a lot of money you literally could
28:43use it for whatever you wanted to use it
28:44for commercial private teaching whatever
28:47just by clicking on it and that project
28:49still exists today in many countries now
28:53high level enthusiasm among participants
28:56so you were obviously participant
28:59one-zero-zero-one of the Personal Genome
29:01Project you're an open book yeah I mean
29:03if you go to your lab website you have
29:05everything you're working on and
29:07everything you've ever worked on you've
29:08described your phenotype in detail which
29:10i think is fascinating did you learn
29:12anything from having access to your own
29:14genome that you found particularly
29:16interesting or enlightening so I didn't
29:18expect to because I felt that I was
29:20likely to be in the 96% that would get a
29:22blank report as it turned out it did
29:25learn a couple of things so one of them
29:28my family was very concerned about
29:30because I had a family history of
29:31cognitive decline was that I had no risk
29:34factors for Alzheimers to this apoe4
29:38static PP precent no one wanted to every
29:41known factors that was reassuring
29:43although I try to tell people not to be
29:46reassured that there's always something
29:48new to learn secondly I'm a alpha
29:511-antitrypsin compound heterozygote
29:53which just means I have two different
29:55risk factors that result in a risk for
29:58lung disease so I should probably avoid
30:00pollution which is probably not a bad
30:02thing for everybody to avoid and smoking
30:05and those were the two main things that
30:08I learned so it's not that different
30:10from getting a blank sheet quite frankly
30:13but we more importantly was having my
30:15medical records publicly available meant
30:18that hematologist gave me personal
30:19advice on my incorrect use of statins so
30:23so it turned out that I was not being
30:26properly diagnosed going back to were
30:27under diagnosed oh wow and I was having
30:29a poor reaction to Staunton as well as
30:32low efficacy it wasn't doing his job and
30:34so we tried a little bit of nudging them
30:36around and finally gave up when I showed
30:38and determined at a vegan diet strict
30:41vegan diet was enough to bring me down
30:43from almost 300 to almost 200 so it's
30:48not generic advice is something that's
30:50very personal and precision and
30:51empirical as I was a there was another
30:54advantage of having people look on and
30:56then there was an advantage to the
30:57project of me being guinea pig number
30:59one the IRB Harvard Medical School IRB
31:03asked me that the IRB is institutional
31:05review board there's
31:06ethics and protocol reviews of human
31:09subjects research they wanted me to
31:13participate as initially the only
31:15subject nor was at least part of the
31:17first ten and that was beneficial in
31:20that when we were developing the skin
31:22biopsy for induced pluripotent stem
31:25cells the skin biopsy first when we
31:27tried on to me was ridiculously pain I
31:30remember that it great and it was in
31:32retrospect was crazy it was like a six
31:34millimeter punch twelve stitches no
31:37anesthetic or at least not in the right
31:40place and then we switched over to a
31:43cream anesthetic which is instead of
31:45twelve injections in the wrong place it
31:47was cream in the right place and then a
31:49simple bandage rather than stitches and
31:51a one millimeter punch so so that was an
31:53example for me being I witness or write
31:57hey I I said no that's not an acceptable
32:00protocol immediately right and I might
32:02not have said that if I were like
32:04detached and I just said to one of the
32:06staff physicians oh just go do it so
32:09that's a summary of why sometimes it's
32:11important for the top researcher to also
32:15be a guinea pig in the study and I don't
32:18think this applies all studies but it
32:20certainly applied to the Personal Genome
32:21Project so switching gears to the church
32:24lab so if you go on your website you
32:26have a list of sort of the active
32:28projects that you're working on and then
32:30it almost reads like screenwriters for
32:32like coming up with the next you know
32:33the great movie talked to me about the
32:35church lab how do you think about what
32:37you work on and even one step before
32:39that how does one get into the church
32:41lab because from an external standpoint
32:43I mean this is like Willy Wonka's
32:45chocolate factory for science so what do
32:47you look for in incoming students for
32:49the church lab and then let's go from
32:51yeah so a lot of it looks like science
32:54fiction and most people would run away
32:56from that not run towards it and they
32:57did when I starting out but now we have
32:59a track record the same level of
33:01creativity and risk-taking but actually
33:05many of the things we do are they look
33:07hard from the outside but from the
33:09inside they look like they're
33:10low-hanging fruit and they happen way
33:12ahead of schedule so for example things
33:14that did look like science fiction were
33:16fluorescent next-generation sequencing
33:19and nanopore sequencing both of those
33:21were wacky when I started them in the
33:241980s and the whole idea that you could
33:26bring down the price of the genome from
33:283 billion down to now sub thousand
33:30dollars also seemed science fiction but
33:34now that we've done it now it becomes a
33:35beacon for people say oh whoever did
33:37that we should go there and if oh if it
33:39the same lab also helped bring in
33:41multiple ways of doing genome editing
33:45if you just do one you could be lucky
33:47but you do several different ways of
33:49doing next-gen sequencing several
33:50different ways of doing editing then
33:53that's an attractant to get in self
33:56selection is another major filter would
33:57do such quirky stuff that people don't
33:59even bother to apply unless they're kind
34:01of already on our wavelength so then the
34:03biggest filter for me and I tell us in
34:05the first interview the first
34:06conversation I have is we're looking for
34:08people that are nice we're not
34:10necessarily looking for geniuses we got
34:13plenty of geniuses we're looking for
34:14people that are nice and how does one
34:16demonstrate niceness well you know I
34:18think some extent just having that
34:19conversation if they want to be
34:21cutthroat they're not gonna come back if
34:22they're kind of sitting on the fence
34:24then they're going to rise to the
34:26occasion they're going to be influenced
34:28by that conversation and by all the
34:31people that have already passed through
34:32that filter that are in the lab and you
34:34create a culture where you try not to
34:38compete with other labs if you can avoid
34:40sometimes it's unavoidable but you can
34:42avoid it by inviting them to work with
34:44them leaving alone fields where there's
34:46plenty of momentum and a lack of
34:48interest in collaboration you're making
34:51sure there's a diverse enough set of
34:52ideas going on in the lab so that
34:53everybody gets to leave with a subset of
34:56those ideas as a parting gift or
34:58continue to collaborate if they want to
35:00as long as they want to so I think you
35:03build up this momentum of knocking off
35:05things that look like science fiction
35:06turning on a science fact
35:08and create a culture of ability to fail
35:12and a bit jump back and to be nice to
35:14your colleagues within and outside the
35:15lab so if I go through the list of
35:17things that you're working on it's a
35:18pretty you know broad array of things so
35:21you are Chris purring dogs to keep them
35:23young you are chris purring pig organs
35:26or had been working on editing pig
35:28organs to make them useful for
35:29transplantation and then you run the
35:32other end of the spectrum you Riaan
35:33in biology to create a mirror universe
35:35of things that would be essentially
35:38immune to all known viruses or microbes
35:39how do you pick the projects what is it
35:43about what's in the water and that one
35:44the Iglesias lab formerly known as a
35:46church lab like what's in the water that
35:48gets this lab to produce so many
35:51startups and spinouts what is that
35:53entrepreneurial energy that sort of been
35:55fostered and created here right so it
35:58may look like a diverse set of projects
36:00but they're actually have common thread
36:03that is surprisingly focused meaning
36:06most people wouldn't fit in this lab
36:08because we're sort of into radical
36:10transformative technologies not
36:12incremental a lot of labs don't even
36:14want to touch technology until it's
36:16working in a company we work years
36:19before that company and we create the
36:22company and then the company has another
36:24few years before us sufficiently worked
36:27out that it can be adopted by a
36:29technology an adoption lab which is
36:32most scientists so anyway that's one
36:35thing that we're a little bit on the
36:36edge and it's an acquired taste or maybe
36:38even a rare taste what's an idea that
36:40was pitched to you that you said wow
36:41that's too crazy well I'm usually the
36:43one pitching the crazy ideas I'm not to
36:45say that we don't have a lot of
36:46creativity in the lab it's pretty rare
36:48that in fact we've banned the word
36:50impossible we certainly try to behave
36:53ethnically but I think that many things
36:55there's a technological solution to some
36:57of the ethical components on all of them
36:59and we try to explore creative solutions
37:02to ethical problems Personal Genome
37:04Project was one of those creative
37:06solutions surveillance for synthetic
37:08biology is another one that I suggest in
37:102004 biocontainment using recoding is a
37:14way that we can make any organism
37:16resistant to all viruses and horizontal
37:18transfer most of these things now work
37:20and many of these things are now
37:22companies that's correct
37:23and their foundation was some sort of
37:25safety ethics component to the company
37:27and part of the secret sauce is hidden
37:30in plain view like you say we're quite
37:31transparent we can keep other people's
37:33secrets but our own we try to get people
37:35to adopt them part of the thing that we
37:37do that is instead of saying failure is
37:40not an option which was one of the
37:41Apollo slogans we say feel fast just
37:43pick yourself up quickly have a bunch of
37:45things going in parallel
37:47the low-hanging fruit empirically as
37:49well as theoretically and just a lot of
37:52things people reject too easily they
37:55either don't think of it at all
37:56or they think about rejected and so if
37:58we see something that looks a little
37:59hard we'll put it up on the shelf orange
38:02plain to view so we can keep reminding
38:04ourselves whenever a new technology
38:06makes that possible we pull it back off
38:08the shelf and we do it and so we have
38:09that culture of constantly re-evaluating
38:12things that are on the edge of science
38:14fiction do you recruit entrepreneurs
38:16that have happened to be scientists or
38:18are you turning scientist into
38:20entrepreneurs I mainly recruit people
38:22who are multilingual multi-disciplinary
38:26because I found this hard to build a
38:28multidisciplinary team from
38:31disciplinarians you have to have a lot
38:33of people who already have done two
38:35things and even if you get two people
38:37have done two things each they don't
38:39have to overlap but they've done enough
38:40translation that they can start talking
38:43to each other and if you have enough of
38:44those multidisciplinary individuals then
38:46you can sprinkle in a few
38:47disciplinarians and you have an amazing
38:49team so the church lab you were pioneers
38:52in sequencing so reading DNA you were
38:54pioneers in CRISPR so writing DNA what
38:59well so there's three-dimensional
39:01structure of living organisms so we'd
39:03really like to know every voxel every
39:05volume element of every pixel in body of
39:08an embryo or a larger section of tissue
39:10we would like to know every molecule
39:11here and we now have tools for doing DNA
39:14RNA and protein in 3d at super
39:16resolution finally that's one thing we
39:19would like to be able to do higher
39:21levels of multiplexing in terms of
39:22editing synthesis of genomes so some
39:24people call it we call this GP right and
39:26her genome project right but it could
39:28just be heavy editing so we set the
39:30record of 62 edits and the pigs and we
39:32now have edit so we have 10,000 edits in
39:35a single cell it's unbelievable so it
39:36goes from 2 to 62 to 10,000 and we want
39:39uses for each of these things so each of
39:41these projects we have a driving
39:42societal benefit for each and we have a
39:46driving technology where we say we don't
39:48just want a factor of 1.5 when a factor
39:51of a million or 10 million and so that's
39:53what pushes each of these projects is
39:54that triple criteria which is cool basic
39:58science philosophically interesting
40:01technological factors of a million and
40:03societal benefit last question ten years
40:06from now or just looking forward to the
40:08future do we get the Neanderthal baby or
40:10do we get the mammoth calf first well we
40:13never really said to in hand with all
40:16baby there's a response to a journalist
40:18whether it was technically possible but
40:20nobody has articulated a reason to do it
40:22but for the mammoth there are lots of
40:24reasons both for the environment and for
40:26enriching the diversity of living
40:28endangered species so we're this is not
40:31about de-extinction this is about making
40:33hybrids and many of these species are
40:35already hybrids of multiple species but
40:38now we can have the benefit of synthetic
40:40genes and ancient genes great well thank
40:43you George thank you for making time and
40:44it's a real pleasure no thank you
