IN 1987, WHEN researchers first used the word genomics to describe the newly developing discipline of mapping DNA, Eric Green had just finished medical school. A few years later, he found himself working on the front lines of the young field’s marquee moon shot: the Human Genome Project. To lead the nation’s participation in the global effort, Congress established the National Human Genomics Research Institute, or NHGRI, in 1989.
Sequencing the entire human genome began the following year, and it took 13 years to complete. Not long after, in 2009, Green took the helm of the research institute. By then, NHGRI’s mission had evolved to include expanding the field of genomics into medicine. That meant funding and coordinating projects aimed at pinpointing the mutations responsible for genetic disorders, then developing tests to diagnose them and therapies to treat them. And even more broadly, it meant generating evidence that DNA data could effectively improve outcomes, even for people who don’t suffer from rare diseases.
To help chart that course, one of Green’s tasks is to periodically put together a strategic vision for the field. Aimed at celebrating progress, identifying technological gaps, and inspiring scientists to pursue the most impactful areas of research, his team published its latest projection in October. For the first time, Green and his colleagues outlined a set of 10 bold predictions about what might be realized in human genomics by the year 2030. Among them: High schoolers will show off genetic analyses at the science fair, and genomic testing at the doctor’s office will become as routine as basic blood work.
Three decades after that sequencing race began, we’ve perhaps reached the end of the early genomics era, a period of explosive technological growth that led to breakthroughs like the sequencing of the first dog, chicken, and cancer cells and the advent of cheap home DNA tests. The field has matured to the point that genomics is nearly ubiquitous in all of biology—from fighting invasive giant hornets to brewing better-tasting beer. Genomic medicine is no longer theoretical. But it’s also not widespread. Although scientists have mapped the human genome, they do not yet completely understand it. Green spoke to WIRED about what the next decade, and the next era in genomics, may have in store. This interview has been edited for length and clarity.
WIRED: October marked the 30th anniversary of the Human Genome Project. When you look around at where we are today, how does it live up to the expectations you had for the impacts the project would make in medicine?
Eric Green: I was inside the Human Genome Project from day one, and I can’t stress enough how back then we didn’t know what we were doing. We had this big audacious goal of reading out the 3 billion letters of the human instruction book, but we didn’t have the technology to do it. We didn’t have the methods. We didn’t even have a functional internet. There was no playbook. So, as someone who got into this as a young physician, I could sort of imagine that one day genomics might be part of clinical care. But I truly did not think it would happen in my lifetime.
If we go back just 10 years, nobody was really using genomics in health care. We fantasized then about the idea of having a patient in front of us, where we did not know what was wrong with them, and being able to sequence their genome and figure it out. That was a hypothetical in 2011. Now it's routine. At least for people suspected of having a rare genetic disease.
That’s amazing. But also, it’s still a far cry from some of the hype around what the Human Genome Project was going to accomplish. In his remarks at the White House in 2000, then-NHGRI director Francis Collins said it would likely take 15 or 20 years to see a “complete transformation in therapeutic medicine,” promising personalized treatments for everything from cancer to mental illness. Obviously, that hasn’t exactly come to pass. Why not?
Part of it is the sheer complexity of genomic information. If physicians were ready to use that information, and patients were ready to act on it, then investing the $1,000 [the going commercial rate] to sequence any of our genomes would be trivial in the grand scheme of our medical care for life. So I don’t think that’s the issue. The issue is that at the moment, for a generally healthy person, we wouldn’t know what to do with that information. That’s why I haven’t had my genome sequenced yet.
No. Because we have the technical ability to generate the sequence, and a very good quality one at that. But then there’s this massive gap between having the data in front of us and knowing what it all means. That’s why one of our bold predictions is to get to a place where we know the biological function of every human gene. We’re making progress, but that progress is likely going to be measured more in decades than in years.
Are there any emerging technologies you can point to that are accelerating progress toward closing that gap?
I need go no further than this year’s Nobel Prize in chemistry: Crispr. A lot of times people hear Crispr and think of therapies for people. But by far the bigger use is at the bench. With Crispr, we can make edits to little pieces of DNA that never go into a person—they go into cell lines or bacteria, which then get tested to see if those edits have functional consequences. The combo of genome editing and genome synthesis methods getting better, coupled with better and better computational tools, is really going to change the pace of biological discovery. Right now we rely on one paper being published about one genomic variant to give us one drip of information at a time. That doesn’t scale.
So we’ve got to get to a point where we’re making millions of changes, generating massive amounts of data, and then hopefully we can use AI to train computers to look for patterns. At that point we won’t even have to do the experiments, because we can make predictions about what a mutation means based on the last 1,000 times we’ve done this. Going forward, those are the sorts of tools that might make the difference.
That sounds like a big lift, in terms of digitizing and analyzing all that biological data.
Of the big challenges that lie ahead for us, at least half of them are computational. It’s a good problem to have. In some ways, we’re the victims of our own success, in that we knocked down so many technical barriers with sequencing that now the big barrier is what to do with all that data. The science has moved so much faster than our ability to plan for some of these things, even in a place like the NIH. If I could wave a magic wand and reorganize the NIH today, there would be a single institute leading in data science. Right now, we don’t have one.
What other barriers do you foresee being a challenge in the decade ahead?
Well, one we’re butting up against right now is that not all insurance companies are willing to pay for a genome sequence. That’s a problem for people with undiagnosed rare diseases. We’ve had much better success in the cancer world, where genetic testing has really gone mainstream, and in prenatal testing. Something like 6 or 7 million pregnant people will get a blood test done this year to screen for fetal genetic defects.
Another one is the uneven uptake of the technology. Patients with rare genetic diseases getting sequenced and diagnosed works really well at Stanford and Harvard and Baylor. But it is not at all working well in rural Montana. So the barrier there is getting doctors who are not at major academic medical centers, who practice in America’s rural heartland, educated and comfortable with genomic medicine. Because the risk we run is the further exacerbation of existing health disparities. If only the richest and most prominent people can get access to genomics, that would be a tragedy. These are the challenges that were once hypothetical, that are now becoming quite real.
How is NHGRI proposing to tackle those challenges?
Well, it’s complicated, of course. These are issues that cut across many aspects of society. But one thing we’re going to be doing in 2021 is unveiling an action agenda for creating a more diverse workforce in genomics—both on the research and the clinical side. If the workforce is more diverse, then genomics will be more uniformly taken up in medicine. So that’s coming.
One of the other projects we’re supporting is an effort to get to a reference genome that captures the full multidimensional diversity of humanity. What we have now doesn’t do that. If we grab someone from the middle of Asia and sequence their genome, we want to compare their variants to an appropriately matched control group so we can assess any rare changes that might be behind a health problem, or contribute to the risk of developing one. If all we have to compare it to is a standard reference that, like the one we have now, happens to be made from European DNA, it can be really misleading. So the goal of this pan-genome effort is to always have available an appropriately ancestrally matched data set available for medical interpretation. Achieving that is also one of our bold predictions.
You mentioned the places where genomics has already become part of mainstream medical care. What corners do you see being the hardest to reach?
The hardest category is going to be preventing common diseases—hypertension, diabetes, cardiovascular disease, asthma, autism, Alzheimer’s, etc. We’re starting to develop polygenic risk scores for these, but we still don’t know how truly predictive they’re going to be.
So these are a way of adding up all the tiny influences of thousands of minute genetic variations, which you can use to estimate someone’s risk of developing these common diseases.
Right. We have major programs investing in big research studies to take polygenic risk scores out for a test drive—to see how predictive they can be and how health care professionals and patients respond to having that kind of information. Because another big question is whether or not they’ll move the needle. If you’re handed a genetic score that tells you you’re at higher risk for becoming hypertensive, for having an early heart attack, say, will that make you watch your diet and exercise and eat less salt? Your doctor might use that information to get you in for an EKG every year starting at age 35, but will you make the appointment and show up? Because that’s the real test—if genomics can actually change people’s behavior.
What about genomics and infectious disease? I’ve written about big efforts, both here and abroad, to mine genetic data to better understand why the coronavirus causes such a wide range of symptoms in different people. How do you see the field contributing to getting us out of this pandemic?
Those major studies are really illustrative of how there’s rarely a problem in biomedicine these days where genomics isn’t somewhere in there playing a role. And they’re going to be really important for helping decipher the extent to which people’s genetic inheritance contributes to their Covid response.
But I think one of the most important legacies of the Human Genome Project was the way it changed forever the way scientists shared genetic data. If you follow the timeline of this pandemic, the first report of the virus was in late December. Within two weeks of that, the sequence of the virus was released publicly.
I remember, that was actually the first story I wrote about the coronavirus—about how it was a real win for public health.
Yes! That sequence was instantly used to make tests for the virus. And it was step one for developing the vaccines that are now being shown to be effective. If you go back to the time before the Human Genome Project, that would have been unheard of. The researchers would have sequenced the virus, written up a paper, submitted it for publication, and a few months later, when the paper came out, they would have released the sequence.
That was the way it was done until we came along and argued that it’s better to give people early access to imperfect data than later access to perfect data. Lots of scientists were worried that would undermine their ability to get credit for stuff. So we also had to bring together journal editors and funders to get them to create and enforce a new etiquette. That was important to us because the Human Genome Project wasn’t a traditional science project. We were creating a community resource. So I think genomics deserves a little bit of credit for changing the cultural norms in some of these other fields, like infectious disease. One of its most lasting legacies is the way it really transformed the rules of research.