Scientist Revisits Data on Raccoon Dogs and Covid, Stressing the Unknowns

A new study of genetic data from a market in Wuhan, China, said the data did not support the case that the pandemic had started with illegally traded animals, touching off fresh debate about samples that other scientists see as critical pieces of the puzzle of how the coronavirus reached humans.

The new study, which examined the relative amounts of animal and viral material in swabs taken from surfaces at the market in early 2020, said it was difficult to draw conclusions about whether given samples of the virus had come from infected live animals or were simply from incidental contamination.

But several outside experts said the analysis, posted online this week by the study’s author, Jesse Bloom, a virologist at the Fred Hutchinson Cancer Center, could have been affected by a number of unknown variables and decisions about how to filter the data.

For those reasons, they said, the findings did not do much to sway their impression of previous studies. Samples from the market containing animal and viral genetic material, they said, were consistent with the possibility that an animal there — perhaps a raccoon dog — had spread the virus to people, but did not prove that had happened.

“I think there’s a pretty reasonable chance they picked up an infected raccoon dog, but that doesn’t prove that was the origin,” said Frederic Bushman, a microbiologist at the University of Pennsylvania who specializes in analyzing samples like those taken from the Wuhan market, but who was not involved in any of the market studies. “I don’t think the Bloom paper changes my thinking that much.”

Chinese researchers wrote about the market data last year and then made the genetic sequences available this year, allowing a team of international scientists to study them. That team wrote in a report last month that based on the data, they could not conclusively identify an animal that had passed the virus to people.

But they said the data confirmed that animals believed to be susceptible to the virus, like raccoon dogs and masked palm civets, a small Asian mammal implicated in the SARS outbreak two decades ago, were being sold at the market in late 2019. Many of the earliest Covid-19 patients also worked or shopped at the market.

Because the market was one of only four places in Wuhan reported to be selling live animals of the sort that could plausibly spread the virus, the scientists said it was unlikely that so many early patients were linked to the market purely by chance. They said the genetic data also built on other evidence, including that two early lineages of the virus had been at the market.

This week’s study took a different approach to analyzing the gene sequences.

Dr. Bloom investigated whether the amount of genetic material from the virus correlated with the amount of genetic material from susceptible animal species in the samples. If one species at the market was overwhelmingly responsible for shedding the virus, he said in an interview, he would have expected to see a clear link between the amount of genetic material from the virus and the amount from that species.

But the study found no clear correlations of that kind. Instead, the strongest correlations involved various fish sold at the market that could not have been infected, an indication that infected people had probably deposited viral material where the fish was.

Dr. Bloom said that finding suggested that the virus, also known as SARS-CoV-2, was spread widely across the market by the time the swabs were collected in early 2020.

“In the same way we shouldn’t read much of anything into the fact that there’s a bunch of SARS-CoV-2 mixed with largemouth bass and catfish samples, we also shouldn’t read much into the fact that there’s a raccoon dog sample with a SARS-CoV-2 read,” Dr. Bloom said.

But outside experts said that various features of the samples could throw off efforts to correlate animal and viral material. The international scientists said in their report that they had considered running a similar analysis, but that it risked producing misleading results. Dr. Bloom acknowledged that “it’s an open question of whether that is an informative thing to calculate at all.”

Genetic material from the virus degrades quickly, said Christopher Mason, a specialist in environmental sampling at Weill Cornell Medicine. Crucially, viral material may decay at a different rate than material from animals, making it difficult to compare them in samples collected over the course of weeks after the market’s closure.

It could be that fish were most closely associated with the virus simply because the fish were likely to have been frozen or refrigerated, slowing the decay of viral material in those samples, said Tom Wenseleers, an evolutionary biologist at KU Leuven in Belgium.

The latest analysis “confirms that looking at these sorts of correlations tells you next to nothing with respect to which host species could have been a plausible source of the pandemic,” Dr. Wenseleers said. This leaves scientists in the same situation as before, he said, with market data that doesn’t offer conclusive evidence of any particular origin scenario.

The new study also looked closely at a swab from a cart at the market in which the international team had found a trace of the virus alongside genetic signatures of raccoon dogs, but no detectable genetic material from humans.

Dr. Bloom wrote that the swab had only a minuscule amount of viral material, and that it was not clear why Chinese researchers had classified the swab as Covid-positive. His study said that swab was the only one that had substantial amounts of raccoon dog genetic material with any traces of the virus.

Some scientists, though, said Dr. Bloom’s analysis risked dismissing other Covid-positive swabs by setting too high of a bar for the amount of animal genetic material in a sample.

Dr. Bushman, of the University of Pennsylvania, said that the threshold used in the analysis was “aggressive” and that it was best to compare results obtained from a series of different cutoffs.

Using a more sensitive threshold, the international team of scientists identified multiple Covid-positive samples containing raccoon dog genetic material, as well as others with genetic signatures of different animals thought to be susceptible to the virus.

Alexander Crits-Christoph, a computational biologist formerly at Johns Hopkins University who helped lead the international team’s analysis, said the team also looked closely at whether the Chinese researchers had been right to describe the swab from the cart as positive for the virus.

He noted that a number of other swabs from the same stall were clearly positive for the virus. He said results from sampling elsewhere in the market also indicated that unlike the swab from the cart, most of the truly negative swabs contained no traces of the virus at all.

“This is environmental sampling of a virus that is a tiny needle in a haystack,” Dr. Crits-Christoph said.