Why Does Bad Science on Covid’s Origin Get Hyped?

You might think that, over the past year or two, public debate about the origins of the pandemic has loosened up. The lab-leak hypothesis has been aired repeatedly in Congress and endorsed with varying degrees of confidence by some intelligence agencies. Even so, pandemic tribalism over whether the virus emerged from a laboratory or in a natural spillover at a market or elsewhere has proved remarkably resilient.

Any shred of new evidence, no matter how flimsy, is hailed as near-definitive proof by one side or the other. In mid-March, new analysis of genetic data collected in Wuhan’s Huanan seafood market from January and February of 2020 enormous publicity for a scrap of virological news — and probably the most attention that’s ever been given to an obscure, adorable species of mammal known as raccoon dogs. The short version is this: An international team of scientists discovered genetic material from the market that was previously held by Chinese scientists and, in analyzing it, found at least one sample that contained, among other things scraped from the market, both raccoon-dog DNA and SARS-CoV-2 viral material.

The analysis was notable, given the absence of hard evidence that an animal had been the source of the pandemic, the over what kinds of animals were in that market and the fact that raccoon dogs had long been seen as a potential intermediate host for the virus. But it was also obviously limited evidence rather than a genetic smoking gun and suggested many follow-up questions. Was there any statistical correlation between the apparent presence of raccoon dogs in certain stalls and the presence of the virus in the same stalls? Did such proximity really indicate an animal infection, or did it simply reflect the presence of SARS-CoV-2 throughout the market? And what could we infer from genetic material collected beginning in January 2020, when we knew already that there were human infections the previous month and possibly earlier? But most commentary about the new report was not nearly so circumspect.

“The Strongest Evidence Yet That an Animal Started the Pandemic,” of the first story to document the analysis. Other accounts were somewhat more equivocal, but readers could be forgiven for skipping past the caveats or for finishing any number of other write-ups and Twitter threads feeling as if a significant and perhaps even definitive piece of evidence had been found. In a few cases, scientists with public platforms in interpreting the data, but in to describe the report as hugely consequential.

For some, the presence of two sets of genetic material in the same sample was the mark of a raccoon-dog infection, and such an infection would be de facto proof of a natural origin of the pandemic. As part of a about the full shape of the pandemic, from origins to the present, Dr. Anthony Fauci highlighted the significance of the new report to me, as well.

Then, last Wednesday, the computational biologist Jesse Bloom presented some persuasive evidence that everyone should perhaps calm down about raccoon dogs. He uploaded a of the same data — this time measuring not just the relative prevalence of animal genetic material but also the relative prevalence of SARS-CoV-2. What he found was quite striking: In the sample that contained the most raccoon-dog genetic material, the presence of the virus was so vanishingly small as to be statistically indistinguishable from zero.

To be precise, in the only sample with abundant raccoon-dog genetic material that contained any SARS-CoV-2 at all, the presence of the virus registered at only one sequence fragment in 200,000,000. Overall, across the full database of genetic material found in the market, the presence of raccoon-dog DNA was negatively correlated with the presence of SARS-CoV-2 material: When samples had more raccoon-dog genetic material, there was actually less SARS-CoV-2 than was found in other samples. “This definitely does not disprove that raccoon dogs could have been infected at some point,” Bloom says.

“But it certainly says these samples don’t provide any evidence of this. ” In fact, the presence of SARS-CoV-2 was much more closely associated with genetic material from other species, Bloom found, notably several kinds of fish and, to a lesser extent, humans. It is implausible, biologically, that the virus came from fish, but Bloom is also careful to say his analysis doesn’t tell us all that much about the relationship with humans, either.

“It just shows that from these data, everything is so mixed up, you can’t reliably make any conclusions about what these associations of animal and viral genetic material mean,” he says. What was briefly hailed as something like a pandemic Rosetta stone turned out to be, on further analysis, something more like a rock with some random scratches on it. This pattern does not apply only to research purporting to show evidence of a natural origin.

Over the past year, we’ve been treated to a series of lab-leak news cycles prompted by and in television interviews, though none have offered any truly significant new evidence. But the pattern is replicated when there is new research to scrutinize and share. Last year, for instance, when a group of scientists — including several who worked on the raccoon-dog report — published of the concentration of SARS-CoV-2 cases at the Huanan market, it was similarly hailed as a major breakthrough pointing to an animal origin from a market spillover.

But when the paper passed through peer review and was finally published, nearly all of its and replaced with more cautious claims, including a statement about the limitations of the study. In the days since Bloom’s new paper was posted, a number of scientists and commentators — a reason to revise our understanding in the direction of an accidental leak. Perhaps, if you staked a lot on that initial raccoon-dog report, it does make sense to turn your dial a bit in the opposite direction.

But more generally, I think, it is a sign of just how much of our understanding of the origins of the pandemic remains shrouded in uncertainty and how terrible we are in accepting that fact. Instead of scrutinizing new evidence, we rush to conclusions at every opportunity, processing all news in deeply motivated ways that blind us to anything that might contradict our preferred understanding and push us to see proof for our own side even when it isn’t really there. How did this happen, and why does it continue to? The answers are perhaps not novel, but they are worth dwelling on.

As Bloom says, a muddled and uncertain story isn’t an especially exciting one, for scientists or for journalists. To many, it can feel that the very value of science hangs in the balance of the pandemic-origins debate, turning it into an increasingly important front in the new culture war. Because the material is so deeply technical, laypeople also impose their own narrative maps on it.

In processing such material, journalists often defer to the analysis of their sources. Across a pandemic in which the public was desperate for new information, we have probably gotten too used to treating hurriedly prepared reports as definitive science. And although the genetic data collected through early 2020 would seem of perhaps limited relevance to the question of what happened in, say, November 2019 — well, that data is about all we’ve got, at the moment at least.

This is all understandable. But that doesn’t mean it’s good. “It is really important to try to understand the origin of Covid-19,” Bloom says.

“But a possible outcome — and certainly the outcome right now — is that we just don’t know. I think part of science, and part of critical thinking in general, is supposed to be a high level of comfort with uncertainty and unknowns. And I think that’s where it really sits right now: This is a situation where there’s such intense interest that it becomes important to say, We don’t really know the answers.

”.