Even in the Facebook/Twitter/Tumblr era, most people don’t put highly personal details like this—especially biological ones, such as medical records and genetic data—out for the world to see. But for Church, the personal is professional. All this openness is part of an experiment he’s running called the Personal Genome Project (PGP), in which volunteers are having their DNA sequenced and making the resulting genotypes, along with relevant details about their lifestyles and medical histories (a.k.a., phenotypes), known to anyone who’s curious.
The point is to provide a preview of a future when many people know what’s in their own genes—a future that will be possible largely because technology invented by Church and others has made genome sequencing much cheaper than it used to be. (In addition to his day job as a Harvard genetics professor, Church advises or has founded more than twenty biotech firms.) The first sequenced human genome cost $3 billion. Today, a sequence can be had for about $10,000, and in the near future, it will cost less than $1,000. This new future will be an astonishingly productive time for medical research, as scientists begin to sift through and compare cheaply obtained genomes to elucidate why some people get sick and others stay healthy. But it will also be a time of shifting ethical and legal norms about biology and identity, and there may be new risks: of genetic information escaping from supposedly secure databases and winding up on the Web, of insurance companies trying to use it against patients, of private data becoming suddenly, irrevocably public.
So Church is trying to figure out what this new society might look like by simulating it in miniature with his band of volunteers
and, in the grand tradition of self-experimentation, himself. All that personal information of his is on the Web because he’s PGP Subject No. 1. “George is the rare practicing scientist who looks outward,” says Misha Angrist, an assistant professor at the Duke Institute for Genome Sciences & Policy, who is PGP Subject No. 4 (yes, his genome and medical data are on the Internet, too). “There aren’t a lot of people developing the technology who are also thinking so far downstream about what will be done with it.”
But let’s go upstream for a moment, back to Church’s time at Duke—because in a sense, that’s where all this started. Church was not an ordinary student. A computer prodigy as a child, he finished his undergraduate degree magna cum laude in two years and went straight into graduate studies. He found himself so fascinated by emerging tools for studying and visualizing molecules that soon he was spending about 100 hours a week in the lab and not nearly enough hours in class. Some of his mentors “tried to argue that I was somehow special,” he says, but administrators drew a hard line: “They said, ‘Yeah, everybody’s special.’ I didn’t particularly hold it against anybody, but I guess I felt like, ‘Well, I’m doing science anyway; what difference does it make if I don’t always go to class?’ That was my immaturity showing.” And so came the letter (“suitable for framing,” Church notes on his website) informing him that he had failed a course and would have to leave campus: “We … hope that whatever problems or circumstances may have contributed to your lack of success inpursuing your chosen field at Duke will not keep you from successful pursuit of a productive career.”
There was no need to worry. Church had already sown the seeds of his productive career during those 100-hour weeks at the bench, developing software for visualizing the structures of tiny molecules of RNA, the complement to DNA that, among other tasks, helps it make protein. (The software is still in use today.) He knew exactly where he wanted to go next: the lab of Walter Gilbert, a Harvard biochemist who would soon win the Nobel Prize. With Gilbert, Church developed a machine that could sequence large amounts of DNA—that is, discern the order of the four chemical “letters” or nucleotides (A, T, G, and C for adenine, thymine, guanine, and cytosine) that make it up.
DNA is an instruction manual for the body’s most fundamental processes. Variants, or tweaks in the order of the DNA letters—a change of an A for a G, say, or the loss of a stretch of letters on a given
chromosome—are a little like typos in that manual. Sometimes they don’t make any difference; the body can still “read” what it’s supposed to do. Other times, the typos result in the manufacture of proteins that don’t perform their biological jobs correctly. Variants can also cause the body to ramp up or dial back the production of many thousands of chemicals crucial to its function. This is how genes cause disease, influence behavior, and, to some degree, make us who we are as individuals. By comparing people’s genetic readouts—especially by lining up lists of hundreds of thousands of genes in many people, side by side, and looking for variants that appear in some people but not in others—scientists can start to figure out why, at the biochemical level, those people differ from each other.
When Church was first developing his technology, this sort of comparison wasn’t feasible. Scientists could read DNA, but
only small stretches of letters—sentences and paragraphs, not chapters, and certainly not entire manuals. It was slow, painstaking work that had to be done by hand. Church and Gilbert’s automating technology (and a similar method called
Sanger sequencing that would eventually overtake it) made it possible to imagine reading DNA on a large scale—to dream not of sentences but of books.
By now it was 1984, and a lot of other people were thinking about the possibilities in the human genome. Among them
were researchers at the U.S. Department of Energy, who that year summoned a small group of scientists—including Church, by far one of the youngest—to Alta, Utah, to discuss the prospect of estimating the rate at which genetic mutations accumulate in the average person under normal circumstances. (They were concerned that the rate would be higher in Japanese civilians who had survived atom-bomb attacks.) Making such an estimate is “barely possible” even today, says Church, because to reach statistical significance requires huge sets of genetic data from thousands and thousands of people, amounts of data that are just now starting to become widely available. It was an unthinkable goal in 1984.
What's inside: Nucleus of each human cell contains forty-six chromosomes (top left); each chromosome consists of a long coil of DNA, which is made of up of base pairs of the four nucleotides: A (adenine) binds with T (thymine), and C (cytosine) binds with G (guanine). If even one base pair is not in the correct sequence in a gene, it can lead to disease. Credit: Christopher Burke
“All of us concluded immediately that there was no way we could estimate mutation rates, so we were basically done with a three-day meeting in the first ten minutes. We said, ‘Hmmm, what else can we do except ski?’ ” says Church. “And then we all kind of simultaneously said, ‘Hey, we could sequence the genome.’ It was like the stupidest meeting I’d ever been to suddenly turned into a smart one.”
The Human Genome Project did not actually involve sequencing one individual’s entire set of chromosomes. Instead, the idea was to make a list of nucleotides that would represent an “average” person, then use it as a sample definition of what the human species looks like biochemically. Scientists could then go on to compare this “reference genome” to those of other, real individuals.
At the time, this was a radical idea. Nothing on its scale had ever been done before in biology. Scientists estimated that it would cost billions, and many worried that it would crowd out other worthy research that also needed funding. But over the next few years, at a series of fractious summits—all of which Church attended—the Human Genome Project crystallized, going from unlikely to inevitable. By 1987, the major mission of biology at the turn of the millennium had been set.
Since then, genomic sequencing has gone from exotic to everyday. The draft of the human genome was released in the early 2000s. (Church helped set up several of the centers that did the work.) Many other organisms, from the cacao plant to boxer dogs to the Tasmanian devil, have been sequenced as well.
Meanwhile, in his own lab, Church has been reinventing the very idea of sequencing. Instead of using the automated techniques he developed in the 1980s, he now conducts research using a much cheaper machine he developed called the Polonator. First, scientists pour a solution of DNA fragments onto a microscope slide. Then they add a polymerase, an enzyme necessary for the replication of DNA. The enzyme causes each fragment to rebuild itself over and over until the slide is covered with millions of miniscule dots called “polonies” (polymerase colonies) teeming with genetic material. These then are “read” by the Polonator, which causes them to light up in different colors depending on which nucleotides—A, C, T, or G—are present
The Polonator is what Church is using to sequence his PGP volunteers. So far, only ten people have undergone analysis and then gone public with the results. All of them are highly educated with some form of expertise in genetics; they include Duke’s Angrist, the technology investor Esther Dyson, and the psychologist/writer Steven Pinker. The roster has a high collective
intelligence quotient by design. Church felt comfortable asking people to reveal their genetic information to the world only if he was sure they understood what they might be getting into. So far, though, all the subjects still have health insurance (a law passed in 2008 forbids health insurers from discriminating on the basis of genetic data). And they seem pretty comfortable with their relative lack of privacy: At a recent conference in Cambridge, they all sat side-by-side on stage, happily taking questions about their experiences so far.
Credit: Christopher Burke
But the Personal Genome Project isn’t stopping with the first ten participants. There’s a broader part of the project that requires a separate group of volunteers to (1) have the protein-making portions of their DNA analyzed, (2) offer up a huge amount of personal information about their health and lifestyles, and (3) sign over permission for Church’s group and other scientists to use all the data for a wide variety of research projects. So far, 15,000 people have signed up. This is stunning, given that the project’s recruiting has been done almost entirely through news articles and word-of-mouth; Church hasn’t been plastering subway cars and telephone poles with fliers. And unlike volunteers in some clinical trials, the 15,000 won’t receive any financial reward. Their only compensation is the information gleaned from their own DNA. This aspect of the PGP alone is unusual. Most genetic research projects don’t give subjects any access to the data that result from the work.
That 15,000 people have expressed interest is even more astonishing considering what the National Institutes of Health thought of the project when Church first proposed it. The agency declined to fund the work, Church says, partly because of “a concern that no one would show up.” (The project is paid for by private donations from individuals and companies such as Google.) Like the Human Genome Project, the Personal Genome Project at its beginning seemed outlandish in both its scope and its open-ended purpose. “George just flummoxed the NIH,” says Angrist. “ ‘Let’s sequence a lot of healthy people and see what happens.’ This was perceived as a totally wacky, beyond-the-pale idea. What bearing did it have on improving health outcomes, curing disease, reducing health disparities?”
The idea of sequencing many healthy people has since caught on; in fact, it’s now the basis for some of the biggest ongoing projects in biology. But Church still has some critics who question the fact that his volunteers have to sign an unusual document, a form based on a principle called “open consent” rather than the “informed consent” typically used in medical research. PGP participants acknowledge that their genetic and health data may be used for purposes that are unforeseeable at present. They are also told that once the project is under way, they’ll be asked to publish their data on the PGP website. Better that than give them a false reassurance of privacy, Church says. “The more people you share data with, the more likely the data will escape. And once it escapes, there are many different mechanisms by which it can be identified.” But some scientists are concerned that this principle of “open consent” could harm the medical and genetic enterprise by scaring people away from trials.
Still, Church has many supporters, among them a young Duke graduate named Daniel Vorhaus ’03, who first read about him in Scientific American back in 2006. “I sent him an e-mail that said, ‘Hey, I’m in law school, I have a background in bioethics, I think this is interesting, is there something I can to do help?’ ” says Vorhaus. “And—classic George—he said, ‘That’s great. Can you write a white paper to NIH defending the project?’ I probably did not have the qualifications to do that at the time. But I was interested, I was motivated, and I was free.”
Vorhaus is now the PGP’s legal adviser, a job he does in his spare time. Since he joined the project, he says, he’s been happy to see how some of Church’s ideas have gone from out-there to almost mainstream. “The notion that people should have access to their own raw genetic information— that used to be an extremely frowned-upon and controversial notion, but it’s been surprising how much acceptance it’s gained in the last couple of years,” says Vorhaus. “Now, people see that as a right. We squabble over what kind of interpretation you can do and what kind of regulation there should be, but the fundamental framework of the conversation has shifted.”
Now that the cultural shift he envisioned and enabled is under way, Church is taking on other projects that are big enough to change society and unusual enough to ruffle a few feathers. There’s his vision for synthetic biology, including the hypothetical construction of “mirror life”—organisms that have the reverse DNA sequences to those found in nature. There’s MAGE, a device that essentially speeds up evolution, inducing billions of mutations in bacterial populations over the course of just a few days, including some mutations that could help scientists engineer bacteria that produce artificial fuels or foods.
And there’s Knome, his personal genomics company, which is offering sequencing services to researchers and
private individuals. It recently picked up a celebrity client, Ozzy Osbourne. Knome’s researchers probably won’t be able to figure out exactly why, despite exposing himself to who knows how many unhealthy substances over the years, Osbourne is still alive. But something interesting is bound to turn up in his DNA nonetheless. Does that project sound a little out-there? Well, yes. It’s classic George.
Duke Magazine’ Newsweek