Print BCTV: Church's Gospel -- Harvard's Church on brain map, open genomes, reviving extinctspecies

Church's Gospel

Transcript of BioCentury This Week TV Episode 132




Dr. George Church, Professor of Genetics, Harvard Medical School, Boston, Mass.



Allen Institute for Brain Science, Seattle, Wash.

Dr. John Donoghue, Professor of Neuroscience, Brown University, Providence, R.I.

National Institute of Standards and Technology (NIST), Gaithersburg, Md.

National Institutes of Health (NIH), Bethesda, Md.

Nature Publishing Group, London, U.K.

U.S. Food and Drug Administration (FDA), Silver Spring, Md.



Steve Usdin, Senior Editor




STEVE USDIN: Creating a brain activity map, building a personal genomics data warehouse, and bringing extinct species back to life -- Harvard's George Church's ambitious agenda. I'm Steve Usdin. Welcome to BioCentury This Week.


NARRATOR: Your trusted source for biotechnology information and analysis -- BioCentury This Week.


STEVE USDIN: Harvard's George Church has been at the center of the genomics revolution and helped launch the Human Genome Project. He developed some of the key technologies that made it possible to sequence the first human genome and pioneered a new generation of faster, cheaper gene sequencers. Now, Church is promoting another bold idea -- an international effort to map the brain in exquisite detail.


The Brain Activity Map project, called BAM, has attracted attention in the White House and among leadership at NIH. The concept hasn't been fully worked out yet, but it's already drawn criticism from neuroscientists who fear it will siphon off funding from other research. Church is also leading another controversial initiative -- the Personal Genome Project. Its goal -- to persuade 100,000 people to publicly share their DNA sequences, medical information, and other personal data. Church isn't shy about breaking down the barriers between science and society. His passion to get scientific innovations out of the lab has led Church to found 16 companies and non-profits and provide scientific advice to scores of biotech companies.


We're pleased to be joined today by this very busy scientist, George Church. He's a professor of genetics at Harvard Medical School and director of the NIH Center for Excellence in Genomic Science. Dr. Church, I want to start with the Brain Activity Map project, what it is, how you go about doing it, and what will be the benefits from doing it. Let's start with the "what is it."


GEORGE CHURCH: So it's still formative, but the objective that my colleagues and I have mapped out so far is to improve the technology and reduce the costs for ongoing neurobiological research of many different types, including those that are already in clinical applications, to be able to measure and manipulate the activity of individual neurons on large circuits of significance.


STEVE USDIN: So the material that's come out, some of the papers that you've coauthored have suggested that the goals in the long run are to be able to look at an entire brain, and to measure the activity, and to be able influence the activities of virtually every neuron in an entire brain. Is that right?


GEORGE CHURCH: I think it may not require an entire brain. We need to be practical about it. But yes, certainly much larger than the number we can do right now. Now we, up until recently, just 100 might be a challenging number of measurements, simultaneously, or manipulations. And so to go to millions, or closer to circuits that include many parts of the brain.


STEVE USDIN: And let's talk about the scale a little bit. How many neurons are there in a human brain?


GEORGE CHURCH: Right. So a human might have 86 billion. A mouse might have 1,000 times fewer -- and as a good model. So you don't have to necessarily do them all, but you need to do a lot more than we're currently doing, which is 100. And you want it to be lower cost, and higher accuracy, and so on.


STEVE USDIN: And what are the technologies that can be used that you're trying to develop that would be able to do that?


GEORGE CHURCH: So there's a number of new technologies coming in, from genetics and synthetic biology and nanotechnology that some neurobiologists are familiar with. But they are very far from being in the mainstream. These include things that are sometimes called optogenetics, where you can read electrical activity in the brain by its influence on things that you can measure with light, and, similarly, you can use light to stimulate the neurons. That's one class. But there are many.


STEVE USDIN: And another class is nanoprobes, right? How would that work?


GEORGE CHURCH: So you can have a, right now, you can stimulate and measure with electrodes which are rather bulky, so it's hard to put a large number, but you can have optical fibers that can go in and can be even smaller -- and electrical ones, or combinations -- they can go in, and you can have arrays of them that are thin enough that they're very, much less perturbing to the tissue around them.


STEVE USDIN: One of the things in some of the papers that you've coauthored -- you've talked about the idea of emergent properties of these brain circuits. That studying them as a whole, you'll learn more than you would buy looking just at individual neurons. What did you mean by that?


GEORGE CHURCH: Well, so many of my colleagues in particular see -- it's just like the sort of things that you might observe in a computer, is an analogy. There's certain things that you could learn by looking at individual parts of the computer. And there are other things that emerge when you look at the entirety of what it's thinking. And in a human, we know that there are circuits that extend over -- that include many different parts of the brain simultaneously.


STEVE USDIN: So the only way that you're going to get public support and funding, probably, for something like this, is to draw kind of a direct connection between the work that you're doing and the potential for advancing the search for cures for diseases. And in some papers you've written you've talked about the potential for this for neuropsychiatric disorders like schizophrenia, for neurodegenerative disorders like Parkinson's. How realistic is that, and what kind of time frames do you think we're talking about?


GEORGE CHURCH: Well, so it is important have it grounded in human biology and clinical applications. And it's not far off at all, in the sense that we already have implants, retinal implants, cochlear implants, for hearing, epilepsy, deep brain stimulation, and even tetraplegics that have brain stem damage -- these are all clinical applications which are generally considered quite crude. They're valuable, but they could be improved.


So right off the bat, we hope to improve those with new technologies while we're developing them for model organisms and so on. John Donoghue, who is one of the early members of the team, has been doing some of this work with direct brain computer interface.


STEVE USDIN: Well, scientists haven't started drafting the Brain Activity Map yet, but other ambitious projects are already under way. Europe is spending $1.3 billion on computer models. NIH is aiming to map the brain's circuits. And philanthropist Paul Allen's Brain Science Institute is providing free access to gene profiles online.




STEVE USDIN: George Church is with us to discuss everything from personal genomics to bringing back extinct animals like woolly mammoths and passenger pigeons. We'll talk about that later.


But first, more about the Brain Activity Map. Dr. Church, one of the things that people have talked about and are concerned about the Brain Activity Map is the idea that it might suck money away from other kinds of research. To start with, what kind of budget do you think is necessary to do the work that you're talking about?


GEORGE CHURCH: So we've emphasized in all of our papers so far that this should be new money that doesn't take away. We actually feel that existing small science -- we want to enable and it's a key part of this.


The kind of budgets that are already in place that we're talking about leveraging, not replacing, is about $500 million, half a billion dollars worldwide, including foundations in the U.S. government, and so forth. That said, what we really want to do is bring down the cost so that each of those small science labs, or medium, can be more effective so they can achieve their goals maybe 1,000, 10,000 times more effectively.


STEVE USDIN: Well, but it sounds like you're talking about two things. One is new money and the other's leveraging existing money.




STEVE USDIN: If you're leveraging existing money, then you are taking money away from existing projects, aren't you?


GEORGE CHURCH: You bring in new money, which then helps the existing money, basically. And I think, just like the Genome Project, it will start with small amounts and ramp up as appropriate. If we're delivering reductions in costs that's helping the existing funding, then maybe the amount should go up a bit.


STEVE USDIN: And then you have any ballpark figures? Because the newspapers have talked about funding figures similar to the amounts that were in the Human Genome Project. Is that we're talking about, hundreds of millions of dollars a year?


GEORGE CHURCH: I think the real breakthroughs occurred after the Human Genome Project was over. A new project called Advanced Sequencing Technology Development kicked in. It was very cost effective. It was around $18 million dollars a year. And it resulted in about a million-fold decrease in the cost and improvements in accuracy and so forth, which has had impact all through biology, not just in genomics. So that million-fold improvement and cost effectiveness that we saw after 2005 has been amazing.


STEVE USDIN: And how would you get that scale improvement in the kinds of neural technologies that we're talking for the Brain Activity Map?


GEORGE CHURCH: Well, in the case of genomics, it was changing the paradigm from the way you measured, from capillary electrophoresis to nano-scale imaging. And it's going to be the same thing here.


There's a lot of nanotechnology and nano-imaging and optical methods that is just ready, is overripe, ready to contribute to the neuroscience community. You can already see it starting to leak in.


STEVE USDIN: And so the other thing -- in biology, the Human Genome Project really kick started the idea of bioinformatics. And here, you're really talking about really ramping up neuroinformatics. And one of the other criticisms that people had is that you're going to generate such enormous quantities of data that people don't have any idea how to visualize it, what to do with it. What would be your response to that?


GEORGE CHURCH: Well, very similar issues come up with every large scale project. And you always rise to the occasion. You get better at collecting the right data set. You don't have to collect everything compulsively. You get good at strategically collecting.


You build computing elements that are closer to the detector, things like field programmable gate arrays and other jargon, and algorithms that scale well and reduce it down to data that we can understand. And you look for the first indications.


And most importantly, because we're not just reading -- we're manipulating as well -- you can test things early on. And that gives you a reality check that you're actually understanding some aspect -- not all aspects, but something -- by testing it.


You can say, OK, this is what happened when a mouse did a certain learning task. If we play that back, do we get the mouse to behave as if it just had that experience. And so there are ways that you can show some understanding way before you have total understanding.


STEVE USDIN: One of the things about also lessons coming from the Human Genome Project is that it's important to avoid hype -- peoples over hyped, the speed, the pace, and the scale of what was going to come out of it. In this Brain Activity Map program, a couple of papers that you coauthored, there's been talk about what the implications of it might be and what the ethical issues are.


One of the ethical issues that one of the papers brings up is the idea of mind control. Is that realistic, a realistic thing to be thinking about? And is that helpful to have that in the discussion?


GEORGE CHURCH: Well, I think the way that we would frame it is not mind control, but brain computer interface. That is to say, if you have a tetraplegic woman who could control a robotic arm, then in a way, it's her mind that's controlling the robotic arm. You're reading her mind, in a certain sense, to give her access.


What we'd like to do is have her be able to control her own arm, which would be a breakthrough. So that's not hype. We have clinical applications now. We'd like to improve them slightly. I think absolutely, we need to avoid over-promising.


STEVE USDIN: For more perspective on the Brain Activity Map, including skepticism from some neuroscientists, visit and download the free article, "Big brain science" from SciBX, published by BioCentury and our partner, Nature Publishing Group.







NARRATOR: Now back to BioCentury This Week.


STEVE USDIN: We're talking with Harvard Medical School's professor of genetics George Church. He's leading an effort to have 100,000 human genomes sequenced and placed in the public domain. Dr. Church, again, you're hoping to have 100,000 people place their genomes in the public domain. What is it that you're hoping will be achieved for science from that, and what will be achieved for the individuals who are contributing?


GEORGE CHURCH: So the initiative is now international. There have been branches that have been established, for example, recently in Canada. And each of these could have its own cohort. We have permission to go to 100,000. There's no particular timeline for that. The idea is to get data more shareable. We have an imperative from the Congress to be able to share biomedical data. And that's great for animals. But we wanted to establish a protocol by which you could safely do it with humans.


And so that the Personal Genome Project is not just about genomes. It's also about environments and traits. Having all these collected on individuals the way that you would, say, in a physician's office five years from now, in the era of precision medicine, you'd have this vast amount of data, big data, on each individual. We want to simulate that now so that people, scientists all over the world, can comment on it and say, tweak this here. Or this is where precision medicine is or isn't working. So you need that, you need to have a whole person, not just a little piece of DNA a here and a little snip of tissue there.


STEVE USDIN: So you've got about 1,000 people now who've had their whole genomes sequenced. I'm wondering, one, are there any examples of people who have learned anything -- I know it's not your goal, but are there instances where people have learned something that's been medically actionable for them. And two, do you think that we are at the point in time now where everybody should have their genome sequenced?


GEORGE CHURCH: So the Personal Genome Project is not intended to be a service, where we interpret people's genome. It's research. But since we're making it public, there inevitably is fairly high quality interpretation that's part of the research that is public. And so people have found out things about themselves. For example, John Lauerman, see, you can mention names because they are identified, some of them, had severe leg pains and dark spots in his retina, which he had kind of pushed aside and ignored. But then he found from his genome sequence that he had a JAK-2 mutation, which is something that affects clotting dysfunction. And so his physician followed up on, and confirmed on it, and recommended a baby aspirin on a regular basis. So there things like that that are actionable.


Your second part of the question is whether I feel we're ready for everybody, or let us say a large subset. There are all kinds of technologies. Almost no technology is ready for everybody.


But I think this is a good time because the cost has come down a million-fold. There are plenty of things that are actionable. And I think most people would just say I don't want to hear about anything that's not actionable. And for many people, that will be nothing. But you don't know that until you look.


It's like why should I get accident insurance or fire insurance? I may never have one. Well, you don't know until you find out. So you could think of this kind of like an insurance policy. But unlike insurance policies, which just mitigate the consequences once they happen, here you can actually influence, you can reduce the probability that something bad will happen, if you happen to have something that's actionable.


STEVE USDIN: We've just got a few seconds left in the segment. A quick question, when people give that kind of data and put it out there, they don't have their names associated with it. Is it realistic for them to expect privacy that ultimately people won't be able to identify who they are from their genetic data?


GEORGE CHURCH: That's sort of the point of our project. Much medical research today promises privacy that really isn't realistic. I mean, data can escape. And then, once it escapes, it could be re-identified. We try to recruit individuals who are OK with that. We don't put their names out, but we tell them, imagine your name's out there. It effectively will be. And many of our volunteers do put their name out there so it is identified. And that's more realistic, I think.


STEVE USDIN: One of the interesting things about your Personal Genomics Project is that you're making it open. And you could make an analogy between that to other things that are open, like Wikipedia. Would that be appropriate, and what do you hope to get from making it open like that?


GEORGE CHURCH: Well, so it is a lot like Wikipedia. That's the intention is, if there's something missing from it, then people can come and patch that gap. As it grows, the whole cohort becomes that much more valuable. Once you have genomes, and microorganisms, and brain activity on this cohort, and so forth, it becomes valuable. And each new team can bring their little piece to it. And it's very hard to do that if it's locked away in a vault somewhere where we're only certain wealthy institutions can participate. And there are geniuses all over the world -- some of them are unaffiliated -- that can really make a big impact.


STEVE USDIN: Have you had any examples of anybody doing anything creative, anything interesting, with the data yet?


GEORGE CHURCH: There's all kinds of things that you can't do with conventionally collected medical data. So for example, the National Institutes of Standard Technology and the Food and Drug Administration, so NIST and FDA, have teamed up to make genomes in a bottle. This is a very serious endeavor but for sharing, for developing new diagnostics, and services, and instruments. You need to all be using the same standard.


STEVE USDIN: What does that mean, genomes in a bottle?


GEORGE CHURCH: What they're doing is, they are providing a standard piece of DNA that everybody's using the same lots so that when they develop a new method, they can standardize and vet it in an international sense. And so ours is one of the few projects where we've properly consented individuals for donating DNA that would be out in the world like that.


Another example are contests or prizes that are given for better interpretation software. So that given a genome, here's the interpretation. If you want to figure that out, you need to have that available in some open way to evaluate. And all this about standards and proving new software that's going to be key to this economy of interpretation really requires that we be open about it because if you have black boxes, then you really don't know whether you're getting the quality that you need.


STEVE USDIN: Do you think it will be common for every baby, when they're born, to have their genome sequenced?


GEORGE CHURCH: So every baby that's born in a hospital in the United States, four million babies a year, is currently tested for up to 40 different genetic symptoms that are treatable. I think that it is becoming close to the right cost that you could do a whole genome. Again, you would withhold everything that isn't actionable. But the actionable list has now grown from 40 to more like 2,700. So very, very soon. I don't know exactly when.


STEVE USDIN: Can genomics technology bring back extinct species and should it? George Church will share his point of view next.


NARRATOR: Now it's 21st year, visit for the most in-depth biotech news and analysis. And visit exclusive free content.




STEVE USDIN: We've been talking with Harvard's George Church about human genomics. But he's been in the news lately for work on animals, especially his efforts to bring extinct species back to life.


Dr. Church, I want to ask you about that. How are you going about trying to bring extinct species back to life?


GEORGE CHURCH: Well, this is really more a matter of habitat preservation and conservancy of ecosystems. Because you have so-called keystone species that are necessary for maintaining an ecosystem and some of them have disappeared relatively recently. And the ecosystem might be decaying in a way that we would really find unpleasant. And so if we could bring back the keystone species, the hypothesis is, and this needs to be developed, that that would be valuable.


STEVE USDIN: Then in the first example, one of the ones you're working on is the passenger pigeon. And most people I think when they think about bringing extinct species back, they imagine that you're going to take some bit of tissue and clone it and create a new species that way. That's not what you're doing, is it?


GEORGE CHURCH: Right. So the passenger pigeons went extinct almost 100 years ago and it happened in a very short period of time. And all we have left are the remains of museum specimens. So we have to take the DNA, put it into the computer, and go from a computer back into a living bird in order to recreate it. So it's much harder than cloning. But it's made possible because of a lot of new research that's being done for human gene therapy for example and genomics and so on.


STEVE USDIN: So as I understand it, you're going to look at the DNA sequence for a passenger pigeon, the extinct animal, compare it to the DNA of a living pigeon and then kind of tweak the DNA of the living pigeon, reintroduce that into the bird and then have them produce offspring that have the new genetic characteristics. Is that right?


GEORGE CHURCH: You want something that has at least enough of the original characteristics so it can retake its place as a keystone species that was involved in maintenance of the forests of North America for millions of years. So you don't necessarily have to bring back every single base pair. But you need to bring back the key ones that are sufficient so that it had the behavior that it had, which was amazing -- of having a flock of billions of birds. It was the largest flock in history.


STEVE USDIN: And then the other species that you've talked a lot about is the woolly mammoth. And what's the idea there?


GEORGE CHURCH: So the idea there is that the ecosystem at risk is the Arctic tundra for Canada and Russia, which is largely uninhabited. But it has almost three times as much carbon at risk as all of the rainforests of the world put together and it's at risk of melting. And the mammoths did various things like knock over trees and punch holes through the insulating snow in the winter so that the freezing temperature could go down and maintain the permafrost. So there's a strong hypothesis in the ecological community that this could be valuable for maintaining all that carbon in its frozen state, rather than releasing methane and carbon dioxide.


STEVE USDIN: Very quickly, we've only got a few more seconds left. Timeline, when could we expect to see an extinct species brought back to life?


GEORGE CHURCH: Well, if you're just going for a few traits, we already have new genome engineering tools like the CRISPR technology that could be kicking in the next two to 10 years.


STEVE USDIN: Very well. Thank you very much and thank you for a fascinating discussion. And thank you for watching.


Remember to share your thoughts about today's show on Twitter. Join the conversation by using the hashtag #BioCenturyTV. I'm Steve Usdin and I'll see you next week.