How I Learned to Stop Worrying And Love the Lab-Leak Theory*

Donald G. McNeil Jr.
19 min readMay 17, 2021


In early spring 2020, I reported an article for The New York Times on which I put the tentative headline: “New Coronavirus Is ‘Clearly Not a Lab Leak,’ Scientists Say.”

It never ran.

For two reasons.

The chief one was that inside the Times, we were sharply divided. My colleagues who cover national security were being assured by their Trump administration sources — albeit anonymously and with no hard evidence — that it was a lab leak and the Chinese were covering it up. We science reporters were hearing from virologists and zoologists — on the record and in great detail — that the odds were overwhelming that it was not a lab leak but an animal spillover.

Frankly, the scientists had more credibility.

The other reason my story never ran was that it was 4,000 words long and full of expressions like “polybasic cleavage site,” “RNA-dependent RNA polymerase gene” and “O-linked glycan shields.” Editors would open it, their eyeballs would bleed, and they would close it and find something else to do.

(Back then, editors blanched even at “spike protein” and “receptor binding domain,” but we’ve all had a crash course in virology this year, haven’t we?)

Although it never ran, others like it did elsewhere. The experts all agreed that the SARS-CoV-2 virus was not a deliberate weaponization of a previously known virus and that it had no obvious signs of lab manipulation (more details below). They noted that blood sampling showed that brief “spillovers” of animal viruses into humans happen often without causing large outbreaks.

Therefore, they argued, the odds were that this was another virus that got lucky, like SARS and MERS and the 2009 pandemic flu: it had dwelled long enough inside a civet or camel or pig or something to infect human-like cells, and then had hit the big city.

For about a year, that was the general wisdom among science writers. The “lab-leak theory” migrated back to the far right where it had started — championed by the folks who brought us Pizzagate, the Plandemic, Kung Flu, Q-Anon, Stop the Steal, and the January 6 Capitol invasion. It was tarred by the fact that everyone backing it seemed to hate not just Democrats and the Chinese Communist Party, but even the Chinese themselves. It spawned racist rumors like “Chinese labs sell their dead experimental animals in food markets.”

China retorting to Trump administration nonsense with nonsense of its own — such as suggesting that U.S. military officers planted the virus during a visit to Wuhan in October 2019 — did not help.

And now to the present day.

Two weeks ago, my former New York Times science news colleague Nicholas Wade wrote an article in the Bulletin of the Atomic Scientists (and on Medium) arguing that the lab-leak theory deserves a harder look.

It has since been sent to me a dozen times with notes asking “What do you think?”

My first reaction was dismissive, even though I very much respect Nick as a journalist. (Some of his work is controversial and he can be cranky, but who am I to criticize anyone on those grounds?) I covered the pandemic from its earliest days and I disagreed with his retelling of how the leak-vs.-spillover debate began.

Also, I was offended by some aspects, such as his attacks on Dr. Anthony S. Fauci of the National Institutes of Health and Peter Daszak of the EcoHealth Alliance, both of whom I have known for years; I know both are dedicated to saving lives, and they have always told me the truth — or what they honestly believed to be the truth at the time, because evidence sometimes changes. They are now both getting death threats, and that is repulsive.

The N.I.H.-funded EcoHealth Alliance does not do dangerous lab research; it doesn’t even have a lab. It hunts for dangerous viruses in the field; its zoologists teach people how to safely gather samples from bats, birds, chimpanzees and other creatures fortified with claws, teeth, beaks, muscles and pathogens.

That’s work I consider as essential as staffing the radar stations that watch for missiles coming over the North Pole. The Trump administration was insane to cut off funding for it. You need to know what’s coming at you. Actually cooking up novel threats is a different matter, of which more below.

I was also bothered by Nick quoting Dr. Robert Redfield, the former director of the Centers for Disease Control and Prevention. To my mind, after being warned about the virus by his Chinese counterpart in the first week of January, Dr. Redfield failed to shout from the rooftops and move mountains — and now 600,000 Americans are dead. He also raised the specter of a flu-Covid “twindemic” that turned out to be virological alarmism.


The deeper I read into the papers and articles Nick cited, the clearer it became how much new information had trickled out in the last year. Not new to the most intense and well-educated followers of this topic, but new to the greater public debate. I include articles like this, this, this, this and this by Yuri Deigin, Rossana Segretto, Milton Leitenberg, Josh Rogin, Nicholson Baker and others.

And more and and more scientists feel misled.

I now agree with Nick’s central conclusion: We still do not know the source of this awful pandemic. We may never know. But the argument that it could have leaked out of the Wuhan Institute of Virology or a sister lab in Wuhan has become considerably stronger than it was a year ago, when the screaming was so loud that it drowned out serious discussion.

And China’s lack of candor is disturbing. It denies access to the institute’s lab logs and whatever messages were swapped during its own investigations, took down 2018 statements critical of lab biosecurity protocols, retaliated against Australia for advocating an open investigation and sharply restricted the W.H.O. investigators.

Calls for a better probe are mounting. Last week, 18 biologists, including leading and outspoken experts on this pandemic like the Harvard epidemiologist Marc Lipsitch and Yale immunologist Akiko Iwasaki, published a letter in Science calling for a new investigation and demanding that Chinese labs and public health agencies open their records to outside scrutiny.

To my mind, China could be forgiven for its standoffishness in early 2020. It was busy fighting its own pandemic. And if China had, say, arrogantly offered to teach the American C.D.C. how to investigate America’s killer hamburgers — the equivalent of the way the Trump administration spoke to China back then — we would have snubbed them too.

But now, 17 months later, China is persistently acting like a nation hiding something.

Also worrying: the hunt for the spillover theory’s smoking gun — a very closely related natural virus in a human or an animal — has gone on for over a year. Success would mean big prizes for the discoverer — especially from the Chinese government, which could say “See??”

And yet — zip. That doesn’t mean it won’t be found. But by now we might have expected at least some smoking shell casings.

I had been skeptical of the “lab leak” theory because animal spillover is such an obvious answer. Genetics has proven that almost every disease mankind has faced jumped from animals: bubonic plague from rodents, measles probably from cows, whooping cough maybe from dogs, and so on.

Also, the leak idea was just too conveniently conspiratorial.

I’ve covered several pandemics and studied others, and one element is consistent: they start in utter confusion that defies any sense that an evil genius is at work. Doctors know something’s wrong, but aren’t sure what. That was true when American veterans started dying of pneumonia after a 1976 convention (Legionnaire’s disease); when the Bronx Zoo’s birds started dropping dead in 1999 (West Nile virus); when young nurses fell ill in Mexico City in 2009 (swine flu); when camel butchers died in Saudi Arabia (MERS); and when Brazilian babies were born with shrunken heads (Zika).

This pandemic’s opening days were also shrouded in fog, and yes, there was a government coverup. But it was outed immediately and it didn’t emanate from Beijing.

In late December 2019, doctors at hospitals near the Huanan Seafood Market began seeing a strange viral pneumonia they couldn’t identify. On Dec. 30, the Wuhan Municipal Health Commission issued a warning.

It was quickly picked up by disease-alert websites like FluTrackers and ProMED; the W.H.O. put out an alert the next day. The New York Times wrote its first article on January 6, from Beijing; I started helping my China colleagues soon after.

By that time, one coverup had been underway for a week. Wuhan’s politically ambitious mayor, Zhou Xianwang, was eager to protect the local party congress he had scheduled for January and the pot-luck dinner for 40,000 Wuhanese he hoped would get him into the Guinness Book of Records.

On January 1, his police silenced the nervous doctors. The Huanan Seafood Market was closed and hosed down.

That was the equivalent of trampling a crime scene. The market’s wild game sellers — who might have had the infected animal, if there was one — scattered. Any live animals or fresh meat probably went to other markets or into the trash. Customers disappeared. The chance to use the market as the hub of a good epidemiological investigation was lost.

At the same time, other events occurred that looked like coverups, but maybe weren’t. As soon as it was clear that the threat was a dangerous new coronavirus, the local health commission and then the national one ordered diagnostic and genetics labs to destroy their samples or surrender them to high-level biosecurity labs. Most labs chose incineration — another crime scene wrecked.

That smacked of coverup, and was treated as such by the Trump administration, but it’s actually standard safety procedure to prevent outbreaks. Our C.D.C. gave the same order in 2014 when it realized that hospital labs had samples from Ebola patients being treated in Dallas and Omaha. “We told the labs in Texas and Nebraska to destroy them or send them to Fort Detrick,” Dr. Pierre E. Rollin, who recently retired from the C.D.C after 26 years of fighting global outbreaks, told me. “You can call that a cover-up, but it was a public health decision.”

During those first days in Wuhan, a major misconception circulated — that the virus did not spread easily between people. The W.H.O. repeated it, so did we. But that was not necessarily deliberate misinformation. With the market closed, the epicenter had scattered a few dozen cases across a city of 11 million. Very few PCR tests existed, and it was the height of flu season. At such times, it’s hard to know who infected whom with what.

Also, I don’t believe the image of China as a Teflon pyramid with Xi Jinping at the apex, the evil emperor who sees every sparrow that falls. It’s like other big countries, even totalitarian ones: messy, with competing scientists and petty bureaucrats. Its flaws often become public despite Beijing’s rigid control of the internet.

Each day back then, the rumors got more bizarre. Some scientists claimed the virus had snake genes. Others said it was part H.I.V., triggering claims that it was a bio-weapon.

Some of that fog of war lifted after Beijing sent Dr. Zhong Nanshan, the country’s renowned epidemiologist, to Wuhan to demand the truth. On January 20th, Dr. Zhong warned on national TV that the virus was spreading rapidly and that outsiders should stay away. On January 23, Wuhan was cut off from the world, Mayor Zhou apologized, and China launched its brutal but amazingly successful effort to crush its epidemic.

The first article I know of blaming the Wuhan Institute of Virology ran on January 26 in the Washington Times, a conservative paper founded by the Unification Church. It seemed based on two elements — the lab was in the same city (albeit nine miles from the market), and a brief, speculative quote from an Israeli biowarfare expert, Dany Shoham.

“Certain laboratories in the institute have probably been engaged, in terms of research and development in Chinese [biological weapons], at least collaterally, yet not as a principal facility of the Chinese BW alignment,” the paper quoted Dr. Shoham as saying. Any work on biological weapons would be “definitely covert,” he added.

When I reached Dr. Shoham by telephone later, he spoke very cautiously. He had not been misquoted, he said, but he emphasized that he had never said that deadly virus came from that lab. He had said only that it “was possible” that such a virus could have come from such a lab.

But the rumor was off and running.

Then, on February 3 — a week later — scientists from that Institute produced what smelled like a smoking gun.

They published an article on Nature’s open-access website saying one of the hundreds of coronaviruses gathered from bat caves that was in their freezers was a 96.2 percent match to SARS-CoV-2.

They called it RaTG13 (indicating a Rhinolophus affinis horseshoe bat captured in Tong Guan cave in Yunnan in 2013).

For conspiracy theorists, that was the clincher — if the lab had a 96 percent match, it must have leaked the killer.

But many of the world’s top virologists leapt to say “Not so fast.”

Coronaviruses mutate slowly, so a 4 percent mismatch in the 30,000 base pairs of the two viruses meant RaTG13 and SARS-CoV-2 had diverged maybe 40 years ago, evolutionary geneticists said.

On February 16, as rumors swirled, five of the world’s top virologists got together to publish a letter on explaining why animal origin was more likely.

The letter, titled “The proximal origin of SARS-CoV-2,” was later republished in the journal Nature.

Its basic argument was that any lab trying to make a super-dangerous virus would start with the backbone of one already known to be pretty dangerous, like 2002 SARS. This new virus was so different from SARS, especially in its receptor binding domain — the crucial bit where the spike protein binds to the ACE2 receptor on the surface of a human cell — that logically no one would have chosen it.

The binding domain was much closer to one that had been recently found in pangolin viruses, so it was likely the pandemic virus had jumped from bats to an animal — perhaps pangolins but not necessarily.

Also, the new virus had a cleavage site unlike those in related coronaviruses. (After binding to a cell, the spike has to “cleave” or split open, to meld with the surface and inject its RNA.) The new virus’s cleavage site was an unusual set of amino acids in an unusual spot on the genome. Such unexpected choices seemed more likely to happen during the constant random evolution that goes on in nature rather than the logic-driven “let’s try this next” methods of a lab.

Also the virus’s spikes had “glycans” which act as shields to protect them from antibodies. A virus sunbathing in a friendly lab cell culture wouldn’t need to evolve shields, while a virus constantly fighting off immune system attacks because it was evolving inside an animal would, they argued.

Therefore, they concluded, “we do not believe that any type of laboratory-based scenario is plausible.”

The paper’s first author, Kristian G. Andersen of the Scripps Research Translational Institute, is still vigorously defending the paper’s basic premise — that an animal origin is by far the most likely.

But other eminent virologists, including at least one of his co-authors, are wavering.

During the uproar that ensued after the institute revealed that the RaTG13 strain was in its freezers, Dr. Daszak argued that merely having a virus in a frozen fecal pellet meant little. Infections take place only when viruses are warmed up and growing in cell cultures or animals.

Dr. Daszak had worked for years with Dr. Shi Zheng-li, the “Bat Woman” who now runs the Wuhan institute. His zoologists and field veterinarians had taught her bat-sampling — a dangerous practice even in caves with tourists in bathing suits wandering around.

Freezers contain hundreds of viruses. It is too expensive to fully sequence all of them and impossible to grow them all out in cell culture, he explained. So labs create a set of “bookmarks.” They sequence one short gene, called RdRP, that seldom mutates and keep a list of their RdRPs. Sometimes they post their lists to a public database like GISAID or Genbank.

Since the world had previously been looking only for relatives of the dangerous 2002 SARS or MERS, he said, labs would care only if a virus’ RdRP gene closely matched those.

“If it doesn’t, it’s of no interest, so you pop it back in the freezer,” he said.

Later, if a new dangerous virus turns up, labs can check their bookmarks for a match, thaw that one out and sequence all of it. That was why RaTG13 was found so quickly, he said. Ditto for pangolin virus sequences found in the freezers of the South China Agricultural University.

Because there were no viruses with closer RdRP matches either in public databanks or in a private Wuhan Institute list of 630 unposted RdRP genes he had seen, he said, the Wuhan lab presumably held nothing closer than a 40-year-distant relative of the killer virus.

“Believe me, if there had been, no one would have kept that a secret,” he said to me more than a year ago. “It would be a huge discovery. We’d be over the moon.”

Dr. Shi herself later told Scientific American that, when news of the new virus erupted, her first fear was that it had come from her institute. She did not sleep for days, she said, until she had finished checking her lab’s logs and assured herself that it had not.

Since then, though, more has come to light about the work done by Dr. Shi’s teams.

The most startling bit of information was that, rather than “finding” RaTG13 in her freezers in February, Dr. Shi had worked with it since at least 2016, but under a different name, RaBtCoV/4991.

RaBtCoV/4991 had not been gathered at random but from a mineshaft in which miners digging bat guano got pneumonia, some fatally. Dr. Shi’s lab sequenced enough of it to be able to say it was the most “SARS-like” of the viruses from that investigation.

There were arguments over whether the miners died of fungal pneumonia, viral pneumonia or both, but that link made it a likely suspect for any lab wanting to explore dangerous viruses. Not mentioning her previous work with it was troubling.

Also, Dr. Shi was trained by Ralph S. Baric of the University of North Carolina in building “chimera” viruses — taking, for example, the spike protein from a new virus and splicing it to the backbone of a known one like SARS. He invented “no-see-um” techniques that left no trace of the splice.

(Interestingly, Dr. Baric is one of the signers of the letter to Science demanding a more thorough investigation of all Wuhan labs.)

Then, to see if the new chimeras could infect people, they were tested against human airway cells and “humanized” mice — mice bred to have human ACE-2 receptors on their organs.

There is debate over whether this is truly “gain of function” research. Some argue that gain of function strictly involves taking a virus already known to endanger humans and trying to make it more lethal or more transmissible.

So Dr. Fauci was answering truthfully in his bitter exchange with Senator Rand Paul on May 11.

But many other scientists feel this is a distinction without a difference. They feel that building any new virus from suspect parts and then seeing if it infects humans is just as risky.

Like nuclear bomb testing, the need for “gain of function” research is hotly contested.

Proponents argue that it is the only way to stay ahead of epidemics: in a world full of emerging diseases, if you can figure out which pathogens are only a few amino acids tweaks shy of disaster, you can develop and stockpile vaccines and antibodies against them.

Opponents say that, noble as that goal may be, it is inherently too dangerous to pursue by building Frankensteins and poking them to see how strong they are.

Despite constantly rising biosafety levels, viruses we already know to be lethal, from smallpox to SARS, have repeatedly broken loose by accident.

Most leaks infect or kill just a few people before they are stopped by isolation and/or vaccination. But not all: scientists now believe that the H1N1 seasonal flu that killed thousands every year from 1977 to 2009 was influenza research gone feral. The strain first appeared in eastern Russia in 1977 and its genes were initially identical to a 1950 strain; that could have happened only if it had been in a freezer for 27 years. It also initially behaved as if it had been deliberately attenuated, or weakened. So scientists suspect it was a Russian effort to make a vaccine against a possible return of the 1918 flu. And then, they theorize, the vaccine virus, insufficiently weakened, began spreading.

Also, Dr. Shi’s teams had done work on inserting cleavage sites into viruses and seeing if that enhanced their ability to infect human cells.

All this raises a disturbing possibility: What if some Wuhan scientist — someone in Dr. Shi’s lab or perhaps at the Wuhan Center for Disease Control right near the market, or possibly some military scientist she trained but could not control — did something like take the likely suspect virus RaBtCoV/4991 and use it as the “backbone” for a set of chimeras with different receptor binding domains? What if that scientist was trying in 2019 to attach binding domains from viruses recently found in dying pangolins seized from wildlife smugglers in southern China? What if someone got tempted to add a cleavage site to see if that supercharged it?

What if various such chimeras were passaged through cultures of human cells or humanized mice? Wouldn’t that speed up mutations into forms likely to infect humans even faster than nature can? Wouldn’t that mean that something that looked like the current pandemic strain could emerge, polybasic cleavage sites, O-linked glycans and all?

And what if someone doing that work in a less secure lab than should have been permitted got infected before catching the subway home?

It’s a lot of ifs, and it’s pure speculation, which has been going on since mid-last year.

Jon Cohen of Science magazine put essentially these very questions to Dr. Shi back in August, 2020.

She said no such work took place in her lab, and that the RaBtCoV/4991 virus had only been sequenced, not isolated or grown out as a virus before the sample was used up. Everyone in her lab had tested negative for antibodies to SARS-like coronaviruses so there was no evidence of an outbreak inside, she said. And she had been assured through regular conversations with other Wuhan labs that that they had no leaks either.

Doubts have been raised about that, including the question: since Covid-19 was racing through Wuhan in early 2020, how likely would it be that no one in her lab tested antibody-positive? Wouldn’t some have gotten infected outside?

Ultimately, much of the debate comes down to this: Is Dr. Shi telling the whole truth? And even if she is, are all her similarly skilled colleagues in Wuhan? Are they being allowed to do so by their government — which has a history of silencing scientists?

Chinese scientists were allowed to interact with W.H.O. investigators only in a very tightly controlled way and very little of the report was devoted to the lab leak theory, which it all but dismissed.

Opening up the 2019 logs of every lab in Wuhan and the 2019–2020 emails between scientists and health officials would go a long way to restoring trust.

And the failure to discover any wild viruses that look like evolutionary intermediate steps on the way to SARS-Cov-2 is troubling.

So virologists are feeling more doubts.

Nick Wade quoted David Baltimore, who won the 1975 Nobel Prize for his work with viruses, as saying the specific amino acid sequences in the cleavage site made “a powerful challenge to the idea of a natural origin.” (This has prompted a complex debate among evolutionary geneticists over which specific rungs on the RNA-DNA ladder are statistically most likely to appear in a bat virus.)

I spoke about Nick’s article last week with Dr. W. Ian Lipkin, the renowned Columbia University virus hunter who was one of the five co-authors on the seminal “proximal origin” paper.

He favored a natural origin theory, he said, in part because he had assumed that all the Wuhan Institute’s 2019 work with SARS-like viruses had been done in its top-level BSL-4 lab, which was cleared to operate in 2017. (State Department cables from 2018 raised questions about how well-run the lab was.)

But later he learned of studies with Dr. Shi’s name on them showing that work he considers dangerous had been done in level BSL-2 labs, which he considers highly porous to leaks, not just in 2016, but in 2020.

“That’s screwed up,” he said. “It shouldn’t have happened. People should not be looking at bat viruses in BSL-2 labs. My view has changed.”

That is still not, as he pointed out, direct evidence of a lab leak. There is no proof of a leak.

But the Occam’s Razor argument — what’s the likeliest explanation, animal or lab? — keeps shifting in the direction of the latter.

The hardest evidence that it was an animal is still what it was early last year: On January 1, right after the market was closed down, and then again on January 12, Huazhong Agricultural University and Dr. Shi’s Institute gathered almost 600 samples from the block-long warren of shuttered stalls.

Of those swabs, about six percent were positive for the virus, according to Xinhua, China’s state news agency. Most came from the western end, where the wildlife was sold. And most, Dr. Shi said, were from spots near or below floor level — the handles of roll down steel shutters and the drains over the floor gutters.

Finding virus in six percent of surface samples was more than might be expected even in a hospital during flu season, Trevor Bedford, an evolutionary geneticist at the Fred Hutchinson Cancer Center in Seattle who does flu studies told me last year.

And the most logical explanation for finding that much virus on the floor and in the drains, he speculated, was not coughing humans. It was the blood of a butchered animal being sloshed around as the market was hosed out.

Yes, there were cases in early December with no connection to that market, but that’s not impossible. Livestock is shipped in batches, Wuhan has other live markets. Also, viruses are known to create “stuttering chains of transmission” as they become more transmissible. We’ve seen the rise of increasingly transmissible variants this year and we know this virus alternates between rare transmission and superspreading.

And the wildlife trade is not some dinky smuggling operation. As the W.H.O. report detailed, there are large farms in China raising civets, badgers and other formerly wild animals for food. A bat virus could have raced through them, adapting itself to more human-like animals, the same way the human virus raced through Dutch mink farms.

Also, farmers all over Asia enter caves to dig bat guano for garden fertilizer. A study Dr. Daszak’s alliance did on villagers living near caves found that three percent had antibodies to bat viruses. That translates to up 7 million inhabitants of rural southeast Asia potentially catching such viruses each year. There may be many small outbreaks that die out without spreading far. Ebola did that at least 19 times we know of between 1976 and 2014, the year the virus reached a big city for the first time.

So there we are. All we have so far is speculation, and all the explanations are unsatisfactory.

The whole thing may just be a cold case, and stay that way forever. But there are more embers left to sift. The whole world, China included, needs a hard answer, whoever is to blame — so we can prevent this from happening again.

*The title is from the 1964 film “Dr. Strangelove or: How I Learned to Stop Worrying and Love The Bomb,” which is about human lies and safety failures — and ends with clips of atomic weapons tests depicted as the real thing. Not that I think the pandemic is funny. But neither is nuclear war.



Donald G. McNeil Jr.

New York Times, 1976–2021. Last beat: lead Covid reporter. 2020 Chancellor Award; 2021 NYT team Pulitzer