It still wasn’t a lab leak

Six years in, all the evidence points in one direction

Early in the COVID pandemic there was understandable confusion over the origin of the virus causing it. The virus, SARS-CoV-2, was similar to the earlier SARS-CoV-1, which had led to illness and death in 2002–2004, but it was also sufficiently different that it wasn’t just SARS-CoV-1 reoccurring. A zoonotic origin¹ seemed likely, but SARS-CoV-2 had not previously been observed in animal hosts and therefore it wasn’t immediately clear what the source was. Some researchers started wondering whether maybe the virus had escaped from a lab.

COVID had a uniquely catalyzing effect on the scientific community. In response to the developing pandemic, many scientists, including both people with a long history of studying infectious disease epidemiology and novices who had never worked in this field, jumped on the opportunity to study COVID origins, spread, and evolution. The field saw a massive influx of people. This prompted me personally to go the opposite direction. While I had published some papers on infectious disease epidemiology prior to COVID, I deliberately stayed away from the fray. For one, I don’t like to work in crowded fields. And also, I’m under no illusion that I can do a better job studying viral phylodynamics² than the likes of Eddie Holmes, Andrew Rambaut, or Michael Worobey. I know when I’m not needed.³

As I watched from a distance how the COVID origin story developed, I noticed a curious pattern. The people that I knew and trusted where all converging on a natural origin, even if they had previously entertained the possibility of a lab leak, while people I had never heard of previously and who had never made a meaningful contribution to infectious disease epidemiology kept shouting “lab leak,” louder and louder. And “lab leak” then morphed into “engineered bioweapon,” a completely ludicrous proposition.⁴

So, the people I knew and trusted were saying one thing, which also made sense to me, and their critics couldn’t even produce a single coherent story of what exactly had happened. Instead they routinely mixed different scenarios or moved goalposts whenever necessary to score an online point.⁵ In this context, it seemed obvious to me which position had the stronger scientific support.

However, it is important from time to time to review what the arguments actually are and what the evidence says. I have done so recently, and here I want to provide some resources that I have found useful. I’m focusing on relatively easy-to-digest material suitable for a broad audience. But I also provide pointers for people who want to dive deeper. Then, below, I’ll close with some additional thoughts about the whole situation.

There are two long interviews on Decoding the Gurus that I’ve found quite informative, featuring the highly respected infectious disease epidemiologists Eddie Holmes, Michael Worobey, and Kristian Andersen. The interviews go into many of the details of the science of COVID origins, including the various pieces of evidence in favor or against natural origins or lab leaks, as well as the types of evidence that could strengthen either case and that may not (yet) be available. The first video is from 2023 and the second is a recent update, from August 2026. If you have limited time, watch the 2023 video, which is quite comprehensive. There is also a good review article by Eddie Holmes from 2024, in case you prefer reading over watching a video or want a direct entry into the research literature. Both of the YouTube videos also list relevant references in their descriptions.

The 2023 video:

The 2026 video:

Now that you’ve hopefully watched at least one of the two videos, here are some additional thoughts, in no particular order.

Tl;dw. Oh, you say the videos were too long and you didn’t watch them? Ok, here are some high-level takeaways. Arguments in favor of a natural origin: The outbreak started at the Huanan market in Wuhan. This market deals with wildlife trade, including animal species susceptible to SARS-CoV-2. We have evidence that such animals were present at the market at the time of the outbreak. Moreover, we now have found SARS-CoV-2-like viruses in both bats and pangolins. So closely related viruses are circulating in the natural world. There is also evidence for continued spillover of these viruses into human populations, so in effect these viruses are constantly testing human defenses and are trying out millions of variants for fitness in humans. A large outbreak is only a matter of time. Arguments against a lab leak: The Wuhan Institute of Virology (WIV), the place where one would expect a lab leak to originate, is not located near the Huanan market, and the outbreak clearly was not centered around the institute. There is also no evidence that WIV had any SARS-CoV-2-like viruses that would have been likely progenitors of the pandemic virus. By contrast, there is ample evidence that the lab experimented with SARS-CoV-1-like sequences.

The scientists followed the evidence, and it did not point towards a lab leak. I believe that anybody who watches the two videos—without any preconceived notions—will come away with the sense that the scientists interviewed had no particular stake in either outcome, natural origin or lab leak, and were just following the evidence. In fact, some of them had originally pointed to the lab leak theory as an important possibility to consider, and then moved away from it as the available evidence simply didn’t support it. The idea that they were coerced or bribed into covering up a lab leak is absurd, in particular for Eddie Holmes who isn’t even in the US and doesn’t depend on US funding.

A lab leak doesn’t imply an engineered bioweapon. One issue that really bothers me in much of the public discourse is the regular conflation of two entirely different concepts, “lab leak” and “engineered bioweapon.” Regardless of the lack of evidence for either in the context of SARS-CoV-2, I want to highlight that these are two distinct concepts and they require different types of evidence.

The term “lab leak” simply implies that the virus was present in a lab and then somehow escaped. This has happened in the past. The most famous example is probably the Russian flu from 1977, where an influenza strain that hadn’t been seen for 20 years suddenly reappeared with few genetic changes. The logical explanation for such an observation is that the strain was stored in a freezer and then leaked as researchers were experimenting with it. Lab leaks happen on occasion and they are documented in the literature.

By contrast, an “engineered bioweapon” would be a virus that was deliberately modified to introduce or enhance properties that would make it more useful as a weapon. Such a virus could then have been released either accidentally or deliberately, so we still have two distinct scenarios we need to consider and evaluate evidence for or against. Demonstrating that a virus was an engineered bioweapon requires a much higher bar of evidence than demonstrating that a natural virus was released accidentally from a lab. It also requires an active bioweapons program, whereas ordinary lab leaks can occur in the course of regular virology research such as vaccine development.

A natural origin is the null hypothesis. The vast majority of viruses infecting humans have their origins in animals and have at some point crossed the species barrier. We have countless recent examples, including influenza virus, HIV, Ebola virus, hantavirus, SARS-CoV-1, West Nile virus, Zika virus, the list goes on and on. Much of the field of infectious disease epidemiology is occupied with studying when and how animal viruses cross over into humans, how to identify outbreaks early, and how to prevent them. Compared to this background of thousands of outbreaks of natural origin, we have only a handful of examples of outbreaks caused by a virus escaped from a lab. And we don’t have a single example of a pandemic caused by an engineered virus.

Therefore, when a novel virus arises and starts spreading and causing human disease, the baseline assumption should be it is of natural origin. That is what happens regularly. Anything else requires strong, supporting evidence, as in the case of the Russian flu of 1977 where no credible virologists doubts it was a lab leak because the evidence is so clear. By contrast, “it is an engineered bioweapon but all the evidence has been covered up” is not a convincing argument. Occam’s razor still applies.

What about the furin cleavage site? Yeah, what about it? Furin cleavage sites arise during natural evolution. There are numerous natural viruses that have them. These sites are on the order of four amino acids long, and several variants work (to varying degrees), so it’s not that difficult for natural selection to stumble onto a working variant. And in particular, the specific furin cleavage site in SARS-CoV-2 very much does not look like it was engineered, as explained in Eddie Holmes’ 2024 review paper.

Closing thoughts. This post contains my personal thoughts on this topic. I am posting it in the hopes that somebody may find it useful, but I don’t want to engage further in back-and-forth over COVID origins. This is my one-and-done contribution to the topic. If you’re not convinced, that’s fine, you’re entitled to your own opinions. Write them up and post them somewhere. I have closed the comments on this article because I don’t want lab-leak conspiracy theorists brigading my discussion board. I consider the matter settled for the time being, though I will probably check back in in five years or so to see where the science stands then. Eddie Holmes promised exciting new data in the 2026 video so I’ll be curious to see how it shakes out.

More from Genes, Minds, Machines

1. “Zoonotic origin” means the virus was originally spreading in some animal population and then jumped over to humans.

2. “Phylodynamics” is the research field studying how disease origins and spread can be inferred from the phylogenetic trees generated in the process of viral transmission and evolution.

3. I also was Department Chair during virtually the entire COVID period, until August 2023, and this was honestly one of the most difficult periods of my life. Research had to take a backseat during this time, and in particular research in a competitive area where it matters whether you publish today or next month.

4. It is ludicrous for multiple reasons, but not least because in 2019, before COVID, nobody would have built something like SARS-CoV-2 if the goal was to develop a bioweapon. SARS-CoV-2 simply didn’t match the preconceived notions of what a bioweapon should look like. See also this somewhat related article of mine.

5. It still boggles my mind that the same people could, with a straight face, both claim that COVID was not a big deal and we all overreacted and also that it was an engineered bioweapon deliberately deployed to kill millions of people. I can’t imagine the cognitive dissonance this must have caused.