Since the SARS-CoV-2 pandemic began three years ago, its origin has been a topic of much scientific — and political — debate. Two main theories exist: The virus spilled over from an animal into people, most likely in a market in Wuhan, China, or the virus came from the Wuhan Institute of Virology and spread due to some type of laboratory accident.

The Wall Street Journal added to that debate this week when they reported that the U.S. Department of Energy has shifted its stance on the origin of COVID. It now concludes, with "low confidence," that the pandemic most likely arose from a laboratory leak in Wuhan, China.

The agency based its conclusion on classified evidence that isn't available to the public. According to the federal government, "low confidence" means "the information used in the analysis is scant, questionable, fragmented, or that solid analytical conclusions cannot be inferred from the information."

And at this point, the U.S. intelligence community still has no consensus about the origin of SARS-CoV-2. Four of the eight intelligence agencies lean toward a natural origin for the virus, with "low confidence," while two of them — the DOE and the Federal Bureau of Investigation — support a lab origin, with the latter having "moderate confidence" about its conclusion.

But it's a theory that the agencies are definitely bringing into the public eye. On Tuesday, FBI director Christopher A. Wray reiterated his belief that COVID-19 "most likely" sprang from "a potential lab incident" in Wuhan, China.

But at the end of the day, the origin of the pandemic is also a scientific question. Virologists who study pandemic origins are much less divided than the U.S. intelligence community. They say there is "very convincing" data and "overwhelming evidence" pointing to an animal origin.

In particular, scientists published two extensive, peer-reviewed papers in Science in July 2022, offering the strongest evidence to date that the COVID-19 pandemic originated in animals at a market in Wuhan, China. Specifically, they conclude that the coronavirus most likely jumped from a caged wild animal into people at the Huanan Seafood Wholesale Market, where a huge COVID-19 outbreak began in December 2019.

Virologist Angela Rasmussen, who contributed to one of the Science papers, says the DOE's "low confident" conclusion doesn't "negate the affirmative evidence for zoonotic [or animal] origin nor do they add any new information in support of lab origin."

"Many other [news] outlets are presenting this as new conclusive proof that the lab origin hypothesis is equally as plausible as the zoonotic origin hypothesis," Rasmussen wrote in an email to NPR, "and that is a misrepresentation of the evidence for either."

So just what is the scientific evidence that the pandemic began at the seafood market?

Neither of the Science papers provide the smoking gun — that is, an animal infected with the SARS-CoV-2 coronavirus at a market.

But they come close. They provide photographic evidence of wild animals such as raccoon dogs and a red fox, which can be infected with and shed SARS-CoV-2, sitting in cages in the market in late 2019. What's more, the caged animals are shown in or near a stall where scientists found SARS-CoV-2 virus on a number of surfaces, including on cages, carts and machines that process animals after they are slaughtered at the market.

The data in the 2022 studies paints an incredibly detailed picture of the early days of the pandemic. Photographic and genetic data pinpoint a specific stall at the market where the coronavirus likely was transmitted from an animal into people. And a genetic analysis estimates the time, within weeks, when not just one but two spillovers occurred. It calculates that the coronavirus jumped into people once in late November or early December and then again few weeks later.

At this exact same time, a huge COVID outbreak occurred at the market. Hundreds of people, working and shopping at the market, were likely infected. That outbreak is the first documented one of the pandemic, and it then spilled over into the community, as one of the Science papers shows.

At the same time, the Chinese Center for Disease Control and Prevention found two variants of the coronavirus inside the market. And an independent study, led by virologists at the University of California, San Diego, suggests these two variants didn't evolve in people, because throughout the entire pandemic, scientists have never detected a variant linking the two together. Altogether, the new studies suggest that, most likely, the two variants evolved inside animals.

Evolutionary biologist Michael Worobey helped lead two of the studies and has been at the forefront of the search for the origins of the pandemic. He has spent his career tracking down the origins of pandemics, including the origin of HIV and the 1918 flu.

Back in May 2021, Worobey signed a letter calling for an investigation into the lab-leak theory. But then, through his own investigation, he quickly found data supporting an animal origin.

When the studies were first published online, NPR spoke to Worobey, who's at the University of Arizona, to understand what the data tells us about the origin of SARS-CoV-2; how, he believes, the data may shift the debate about the lab-leak theory; and the significance of photos taken five years before the pandemic. Here are key points from the conversation, which has been edited for clarity and length.

Live animals that are susceptible to COVID-19 were in the market in December 2019

It's clear-cut these wild, live animals, including raccoon dogs and red foxes, were in the market. We have photographic evidence from December 2019. A concerned customer evidently took these photos and videos of the market on Dec. 3 and posted them on Weibo [because it was illegal to sell certain live animals]. The photos were promptly scrubbed. But a CNN reporter had communicated directly with the person who took the photos. I was able to get in touch with this reporter, and they passed on those photos from the source. So we don't completely verify the photos.

Live susceptible animals were held in a stall where SARS-CoV-2 was later detected on a machine that processed animals in the market

We analyzed a leaked report from the Chinese CDC detailing the results of this environmental sampling. Virtually all of the findings in the report matched what was in the World Health Organization's report. But there was some extra information in the leaked report. For example, there was information not just on which stalls had virus in them — or had samples positive for SARS-CoV-2 — but also how many samples in a given stall yielded positive results.

We found out that one stall actually had five positive samples — five surfaces in that stall had virus on them. And even better, in that particular stall, the samples were very animal-y. For example, scientists found virus on a feather/hair remover, a cart of the sort that we see in photographs that are used for transporting cages and, best of all, a metal cage in a back room.

So now we know that when the national public health authorities shut down the market and then sampled the surfaces there, one of the surfaces positive for SARS-CoV-2 was a metal cage in a back room.

What's even weirder — it turns out that one of the co-authors of the study, Eddie Holmes, had been taken to the Huanan market several years before the pandemic and shown raccoon dogs in one of the stalls. He was told, "This is the kind of place that has the ingredients for cross-species transmission of dangerous pathogens."

So he clicks photos of the raccoon dogs. In one photo, the raccoon dogs are in a cage stacked on top of a cage with some birds in it.

And at the end of our sleuth work, we checked the GPS coordinates on his camera, and we find that he took the photo at the same stall, where five samples tested positive for SARS-CoV-2.

So we connected all sorts of bizarre kinds of data. Together the data are telling a strong story.

Earliest known cases of COVID-19, even those not directly related to individuals who had been in the market, radiate out from the market

With a virus, such as SARS-CoV-2, that causes no symptoms or mild symptoms in most people, you don't have any chance of linking all the early cases to the site where the outbreak started. Because the virus is going to quickly spread to people outside of wherever it started.

And yet, from the clinical observations in Wuhan, around half of the earliest known COVID cases were people directly linked to the seafood market. And the other cases, which aren't linked through epidemiological data, have an even closer geographical association to the market. That's what we show in our paper.

It's absurd how strong the geographical association is [to the market].

NPR: Absurd? How? In the sense that the seafood market is so clearly bull's-eye center of this outbreak?

Yes. How could anyone not be moved, at least somewhat, by that data and then not take this idea [of an animal origin] seriously, especially given the other things we've found in these studies? That I don't understand.

The virus jumped into people right before the outbreak in the market

For example, our new genetic analysis tells us that this virus was not around for very long when the cases occurred at the market. For example, the earliest known patient at the market had an onset of symptoms on Dec. 10, 2019. And we can estimate that at that point in time, there were only about 10 people infected with the virus in the world and probably fewer than 70.

So if the pandemic didn't start at the market, one of the first five or 10 people infected in the world was at the market. And how do you explain that?

You have to remember: Wuhan is a city of 11 million people. And the Huanan market is only 1 of 4 places in Wuhan that sold live animals susceptible to SARS-CoV-2, such as raccoon dogs.

It's highly unlikely that the first COVID-19 outbreak would occur at the market if there weren't a source of the virus there

Step back and think, "Where is the first cluster of a new respiratory infection going to appear in this city?" It could appear at a market. But it could also appear at a school, a university or a meatpacking plant.

NPR: Or a biotech conference?

Yes. In Washington state, SARS-CoV-2 first appeared in a man who had traveled back from China. In Germany, it was at an auto-parts supplier.

There are thousands, perhaps 10,000, other places at least as likely, or even more likely, to be the place where a new pathogen shows up. And yet, in Wuhan, the first cluster of cases happens to be one of the four places that sells live animals, out of 10,000 other places. If you're not surprised by that, then I don't think you're understanding the unlikelihood that that presents.

NPR: So what is the likelihood of that coincidence happening — that the first cluster of cases occurs at a market that sells animals known to be susceptible to SARS-CoV-2, but the virus didn't actually come from the market?

I would put the odds at 1 in 10,000. But it's interesting. We do have one analysis where we show essentially that the chance of having this pattern of cases [clustered around the market] is 1 in 10 million [if the market isn't a source of the virus]. We consider that strong evidence in science.

The analyses that we've done are telling a very strong story.

The evidence is amongst the best we have for any emerging virus.

NPR: Really?

It's important to note we haven't found a related virus from the intermediate host. But we have a bunch of other evidence.

And the data zeroing in on the Huanan market, to me, is as compelling as the data that indicated to John Snow that the water pump was poisoning people who used it. [John Snow was a doctor in London who helped launch the field of outbreak investigations by figuring out the source of a cholera outbreak in the city in the mid-19th century.]

Making these findings brought tears

Sometimes you have these rare moments where you're maybe the only person on Earth who has access to this kind of crucial information. As I just started to figure out that there were more cases around the market than you can expect randomly — I felt that way. And no exaggeration, that moment — those kinds of moments — bring a tear to your eye.