Before the pandemic, Trevor Bedford was best known in a small circle of bioinformatics specialists who use rapid genomic analysis to monitor pathogens like the Ebola virus as they evolve and spread.
Bedford and his colleagues at Fred Hutchinson Cancer Research Center and the virus-tracking project called Nextstrain were perfectly positioned to serve as a kind of central, scientific command when the novel coronavirus emerged, documenting the tendrils of contagion that followed air and transit corridors, and noting every mutation and genetic quirk along the way.
What had been a little-known field where computer technology and genetics intersected was suddenly a matter of global urgency.
Bedford’s Twitter feed, which now has nearly 250,000 followers, has become a must-read for infectious disease experts and armchair epidemiologists. Health officials turn to the computational biologist and his colleagues for insight and analysis. When genetic sequencing of the first two cases in Washington state suggested the virus had been spreading silently through the community for six weeks and was poised for exponential growth, Bedford sounded the alarm Feb. 29 — via a long Twitter thread spelling out his reasoning — and helped galvanize the public health response.
The quick action is generally credited with preventing the kind of runaway epidemic that devastated New York.
But last week, with new evidence from a thousand additional genomes and a reanalysis by other researchers, Bedford revised that early narrative of Washington’s outbreak.
Instead of originating with a Snohomish County man who returned from Wuhan on Jan. 15 and is believed to be the first U.S. case, Bedford now says the local epidemic was probably sparked around Feb. 1, when a second strain from China made its way to the state. But his analysis still suggests the virus spread undetected for several weeks, infecting about a thousand people by early March and planting the seeds of the epidemic that soon unfolded.
The new details about the timing and viral strains wouldn’t have significantly altered the public health response, even if they had been known earlier, Bedford said in an interview Friday. “I think the response was exactly the right response,” he said. “It worked out for the best.”
The evolutionary picture could shift again as more evidence accumulates — and the entire story of how the virus arrived will probably never be known, he said in a conversation that also touched on the challenges of doing science at lightning speed, the limitations of test-and-trace, and the role of digital technology.
Here’s an excerpt, edited for brevity and clarity:
Q: Science is moving so fast, and a lot of it is being presented on Twitter. Do you regret immediately tweeting out your original theory about the outbreak, now that you know it was incorrect in some regards?
A: No, I don’t regret it. I think it was the best hypothesis at the time, given the data at the time. Now we have new data, so that’s changed. Normally in science that wouldn’t have been the right thing to do, because it was spotty data on Feb. 29 — but the consequences of not saying anything were huge. I think I’ve been more willing to go out on a limb than many of my colleagues and I’m OK with that. Scientists by nature are a careful bunch. But right now, we’re combating a fast-moving pandemic, and in those conditions, you do the best job you can with the data that’s available to inform public health.
Q: What do you know now about the mix of viral strains in Washington?
A: Most of them, 85%, are related to this WA2 outbreak virus (the second introduction in early February), which is quite similar to WA1 (the strain carried by the Snohomish County man, which seems to have been stamped out). There’s another 10% that are clearly imports from New York City and environs that are of European lineage and show up later in March. The other 5% are kind of one-off things that are quite different from anything else and appear to be separate introductions. But when they did land, they didn’t really spark.
Q: You describe the fight against the pandemic as the Apollo program of our times. How do you think we’re doing?
A: Not great. I’d love for there to be a national test, trace, isolate strategy that we could be investing in and ramping up. But we’re left with different states trying to figure it out for themselves.
The thing that is not so easy right now is that we believe we have so much presymptomatic or asymptomatic transmission. What we really need is a cycle where there’s a primary index case that gets tested right after symptoms start, and then their as-yet-asymptomatic contacts would get immediately tested as well. We need to have that cycle happening quickly across the U.S., but it’s not dialed in yet. It’s getting better, but it feels like the urgency isn’t there anymore.
Q: So far, most of the epidemiological investigation and contact tracking by health officials relies on old-fashioned phone calls. You’ve advocated the use of digital tracing technology and Google and Apple have developed apps people can voluntarily opt in for. What role do you see for that technology going forward?
A: With so much social distancing going on, people are interacting with fewer other people in general, so it’s easier to do that kind of telephone interview. As things come back online, the ability to have digital contact tracing could improve things by adding timeliness to the exposure notification. So you’re not waiting to get a phone call (and you can also be notified of) casual contacts, like people in a crowded bus together.
What’s really important now, and is not very well understood, is whether even if you had a lot of adoption of the Google/Apple sort of proximity app approach, how many more actual exposures that result in infection does that data catch, compared to the traditional approach? To my knowledge that’s a completely open question.
Q: What would you like people to know as communities begin to open up?
A: What we’re in at this very moment (is a process of) trying to figure out the more fine-grained knobs to turn in terms of what really limits transmission. Part of it is figuring out: How much is the weather really improving things? And how much can we start bringing things back online while limiting the spread. There’s a lot to learn, like figuring out exactly what are the kinds of settings that should be restricted, and what are the settings that can be opened up. And then moving to more targeted sorts of control measures, rather than being so broad.
Q: You’re a scientist used to working behind the scenes. What’s it like to find yourself as a major influencer in a global pandemic?
A: I did not expect this. It’s mostly obscene pressure that I have a position now to influence things, and I should be working tirelessly … to have the science actually make a difference. And terror, in terms of upholding that end of the bargain.
Previous positions/fellowships: University of Michigan Department of Ecology and Evolutionary Biology; University of Edinburgh Institute of Evolutionary Biology
Honors: Winner, 2017 Open Science Prize for NextStrain, which uses shared genomic data to track epidemics as they unfold. “Open Science Prize”: https://jamanetwork.com/journals/jama/fullarticle/2620070
Education: Ph.D. in biology, Harvard University