A vehicle belonging to the Wuhan Hygiene Emergency Response Team drives past the shuttered Huanan Seafood Wholesale Market in Wuhan, China, on Jan. 11, 2020 — the day that the Wuhan health commission reported what is believed to be the first death from COVID-19. A new study points to raccoon dogs sold in the market as the likely source of the spillover of the SARS-CoV-2 virus from animals to humans.
Noel Celis/AFP via Getty Images
Remember the raccoon dog? The adorable fox-like critter made headlines in the spring of 2023, when it was implicated in some preliminary research on the source of the virus that causes COVID-19. Researchers had found genetic evidence that raccoon dogs, an exotic species known to be susceptible to the virus, were among the animals for sale at the wet market in Wuhan, China, where the outbreak was first identified.
Now that research has been updated and published in the peer-reviewed journal Cell. The authors, a team of scientists that includes prominent names such as Michael Worobey and Angela Rasmussen, say the analysis shows with unprecedented granularity — down to the individual market stall — that coronavirus-susceptible wildlife and the SARS CoV-2 virus were mingling, along with human beings, in a very specific part of the Wuhan market.
Exotic wildlife for sale at the southwest corner of the Wuhan wet market.
Eddie Holmes, University of Sydney, Australia
However, the analysis relies on an imperfect set of data gathered by Chinese scientists, which has left this research open to criticism by some researchers — in particular, those who argue a leak from the Wuhan Institute of Virology is a more likely source of the virus’s spread.
But the new study frames its conclusions as part of a growing body of data showing that the virus spilled over from wild animals to humans at the market, with multiple lines of evidence that all tell the same story.
“This doesn’t prove by any means that there were infected animals at the market, but we believe that to be by far the most likely hypothesis,” says Kristian Andersen, director of infectious disease genomics at the Scripps Research Institute and one of the authors of the new study. “Once these data were made available [by the Chinese government in 2022], it allowed us to very specifically test that hypothesis. And, indeed, we see exactly what we would expect in a scenario in which the emergence of this virus came from infected wildlife.”
Raccoon dogs held in cages for sale at Stall A in the Wuhan wet market. The photo is from Oct. 29, 2014.
Eddie Holmes, University of Sydney, Australia
A “hot spot” in Stall A
In the southwest corner of the market, a handful of stalls were known as the place to buy live, exotic wildlife: Himalayan marmots, hoary bamboo rats, Amur hedgehogs and raccoon dogs, a small cousin of the fox native to east Asia, recognizable by its raccoon-like face, but also bred for fur on ranches closer to Wuhan, in central China. Photos of what would later be known as “Stall A,” taken by Australian scientist Eddie Holmes (also an author on the Cell paper) in October 2014, show wire cages lining the tile floors. In the dim space decked with red and yellow signs, two cages stacked on top of a drain grate house live raccoon dogs
Five years later, amid a mounting and mysterious viral outbreak in Wuhan, the Chinese Center for Disease Control and Prevention would comb through this stall and many others at the Wuhan market, swabbing surfaces to test for genetic material on Jan. 1 and Jan. 12, 2020. This was just after authorities shut down the market because of concerns about the emerging outbreak. In Stall A, the team gathered samples from now-empty cages as well as from the floor, a cart and the drain outside the stall.
The results of that sampling would arrive after a delayed and circuitous route — it was posted on a public server briefly and then later published — into the hands of Western scientists. When Andersen and his colleagues parsed the genetic sequences of thousands of organisms picked up on those swabs from Stall A, they found the genetic footprints of a host of exotic animals — masked palm civets, Himalayan marmots, Malayan porcupines and raccoon dogs — as well as the viral RNA of SARS-CoV-2, the coronavirus that causes COVID-19.
Across the market, they found other locations with a similar stew of genetic markers (though none with a signal quite as strong as Stall A’s). These “hot spots,” as they’re characterized in the study, tended to cluster around the southwest corner of the market, where wild animals were known to be sold.
From this circumstantial evidence, it’s difficult to tease out the explanation for the overlapping genetic evidence. One possibility, which the authors say is by far the most likely, is that animals infected with SARS-CoV-2 were brought into the market and the virus then jumped to humans, perhaps in two spillover events. Another explanation is that humans at the market were already infected from an unidentified source and that they were passing the virus around and shed some of it in the wildlife stalls where they worked or shopped.
But to the authors of the paper, the takeaway is clear. Michael Worobey, head of the Department of Ecology and Evolutionary Biology at the University of Arizona and a co-author of the new paper, says if the COVID-19 pandemic did not originate at the market, all that genetic material turning up where it did would be a remarkable coincidence.
“It doesn’t 100% prove that those animals had SARS-CoV-2, but it shows that you can just say goodbye to the idea that these [coronavirus-susceptible] animals weren’t even there at the time the pandemic started. The ghosts of their DNA and RNA were certainly there. [The animals] were observed there in November by Chinese colleagues. And it becomes yet another piece of evidence that you can throw against any hypothesis for the origin,” Worobey says.
Worobey says the genetic data, taken at least six weeks after the outbreak is thought to have begun, gives a valuable albeit incomplete snapshot of the Huanan market at a crucial moment.
Skeptics of the new research say the data is far more flawed than that — perhaps fatally.
A crucial, and imperfect, set of data
Jamie Metzl’s objections to the new paper start with the underlying data — the genetic information from all those swabs gathered in January 2020. Metzl, a fellow at the Atlantic Council and the author of books on genetics and technology, has been an outspoken critic of the market-origin hypothesis.
The swabs from the market, Metzl says, are not a representative sample but rather a biased one: The Chinese scientists were preferentially collecting data from the west side of the market, where live animals were known to be sold.
“What they did is to say we’re going to prioritize our sampling in a specific section of the market. It’s circular logic to say, well, we have more positive samples concentrated on the western side of the market when, according to the Chinese sources, that was where they were doing most of their sampling.”
Experts in Metzl’s camp point to an analysis of the market data by Jesse Bloom, an evolutionary biologist at the Fred Hutchinson Cancer Center in Seattle. That report, published last year, found that no particular mammal species correlated with the presence of SARS-CoV-2, as one would expect if, say, raccoon dogs were the main viral reservoir.
Metzl also questions the paper’s proposed time line for the outbreak’s beginnings in mid-November as well as its hypothesis that SARS-CoV-2 spilled over from animals to humans in the market not once, but twice — based on finding different versions of the virus. And he takes exception with how the authors characterize the state of the science.
“These same authors are once again trying to promote a false consensus about science that is at very least debated,” Metzl says. “It’s just not the case, as these guys would have you believe, that this is science and the other side are a bunch of crazy people with agendas. This is disputed science.”
A genetic time line pointing to the market?
The authors of the new paper acknowledge the data set’s limitations. They say they statistically adjusted for sampling bias and still found signals of SARS-CoV-2 clustered around the suspect animals. And they say other lines of evidence all align to tell the same story.
For example, the researchers used the slightly different lineages of the virus in the market samples to model its evolution and compare that with a similar analysis of the virus from the broader pandemic. What they found, says Andersen, is that the two origin stories produced by those computer models both point to the outbreak starting around mid-November at the market.
These viral family trees “perfectly overlap between the market and the larger pandemic as a whole, which is another example of this hypothesis-testing framework: This is exactly what we would expect in a scenario in which all of this started at the market itself,” Andersen says.
Other research has placed the beginnings of the outbreak earlier, in October or even September, which would complicate the market-origin hypothesis. But that theory, too, has gaps. If the virus originated outside the market and had spread under the radar for months, why were there no reports of sick people before December? What’s more, early cases and hospitalizations in Wuhan clustered near the market, even among people with no known connection to the market itself.
“That is the biggest tell that this market wasn’t just somewhere the virus eventually got to after spreading widely in Wuhan. There’s very clear indications that that’s where the jump successfully took place and human-to-human spread began,” says Worobey.
An incomplete picture
Worobey says for all its uncertainty and imperfect data, the new analysis tells a remarkably detailed story of COVID-19’s likely origins.
“Yes, we did not have video rolling of a civet coughing on a stall owner and samples from the nasal passage of each, which we never do for any pandemic or pathogen,” Worobey says.
“But we now have, with these genetic data and multiple other lines of evidence, in many ways, a much richer record of what happened in this case than we do for HIV or the Spanish flu or even the 2009 H1N1 influenza epidemic.”
What the new study does not show is that any given animal at the market was infected with SARS-CoV-2, meaning it cannot rule out that all the viral samples collected came from humans. Such evidence is unlikely to emerge at this point, given that the market had already been emptied and the wildlife removed before the swabs were collected.
Worobey calls this “absence of evidence” rather than “evidence of absence.”
And he says the latest round of data lines up not just with a hypothesis that the pandemic began at this market but also with the worst-case scenario that virologists and public health experts have been warning about for decades.
“The most likely origin of this pandemic is what we’ve always been concerned about in this field, which is taking these sorts of wild animals that harbor exotic viruses and placing them in the middle of the tinderbox of a big city.”
