The argument continues about whether the virus causing covid originated in a wet market in Wuhan or as an accidental release from The Wuhan institute of Virology. While several U.S. government agencies have agreed that the evidence is tilted towards a lab-leak origin, in my view the evidence is not dispositive on either side.
Matt Ridley, however, has been a hard-core advocate of the lab-leak theory, and even co-wrote a book with Alina Chan that, at the time, presented both sides and, as Ridley says below, he “remained unsure what happened at that stage.”
No longer. Since 2021, Ridley has promoted the lab-leak theory, which he does in a Torygraph article shown below (click on headline below to get the archived version). Apparently Ridley teamed up with another collaborator, P. Anton van der Merwe, and wrote a scientific paper laying out his evidence for a lab-leak origin of covid. I’ve put the paper’s title below, but you can read it at the same Torygraph site. The scientific argument was published in the newspaper rather than in a scientific journal because the journal rejected it. (No explanation is given.)
In the intro before he shows the paper (surely a first for the Torygraph), Ridley explains how this came about:
In 2024 I was approached by a single member of the editorial board of a respected biological journal with a request that I team up with a British biologist with relevant expertise and compose an academic paper setting out the case for the lab leak hypothesis: he hoped the journal would consider it. With the help of Anton van der Merwe of Oxford University, and advice from Alina Chan, I drafted such a paper. The paper was rejected; I suspect that it was another case of not wanting to rock the scientific boat.
Now I am posting this paper online for all to read. It was composed several months ago so one or two small new items may be missing, but nothing in it has proved wrong. It is written not in my normal style but in dry, scientific prose, with each statement backed up by a source, in the shape of nearly 100 end-note references, so that readers can check for themselves that we have represented the sources faithfully. It deserves to be available to people to read.
Here is some of the evidence Ridley and van der Merwe adduce:
- Attempts to find evidence for a wet-market leak have been unsuccessful. The cases found around the wet market could simply reflect sampling bias, as the Chinese concentrated on looking for infected people in that area.
- The Chinese have not been forthcoming with their data, and in fact locked one site with a catalogue of the sequenced but unpublished viruses they were working on
- If a person got infected with a bat virus from Yunnan (one theory), that person would have infected others on his/her journey to Wuhan, but there is no such trail of infection
- The Wuhan institute was doing “gain of function” experiments to increase the infectivity of SARS viruses (not the progenitor of the covid virus), but these did involve making viruses more transmissible.
- There were plans to put “furin cleavage sites” into SARS viruses, sites that make it easier for the viruses spike protein to get into cells. The virus causing covid has such a site—12 nucleotides long— which Ridley and van der Merwe insist was inserted into the virus progenitor by humans. As Ridley notes:
When the pandemic began in January 2020, Shi Zhengli of the WIV published two articles, one co-authored with Shibo Jiang, yet in both of them failed to mention the furin cleavage site, by far the most remarkable feature of the new virus’s genome. This may have been an oversight, but by contrast, it was the furin cleavage site that immediately alarmed several western virologists on first seeing the genome of the virus and led to the drafting of the Proximal Origin paper. Messages released during a congressional investigation reveal that the authors of the paper were not themselves convinced that a laboratory origin could be ruled out, either during or after the writing of the paper
- The containment of viruses at Wuhan for the SARS experiments was Level 2, which American scientists think was far too lax for such potentially dangerous experiments (this itself, of course, is not great evidence for a lab leak).
Here’s Ridley and van der Merwe’s conclusion:
In only one city in the world were sarbecoviruses subject to gain-of-function experiments on a large scale involving human airway cells and humanised mice at inappropriate safety levels: Wuhan. At only one time in history was research to create novel sarbecoviruses with enhanced infectivity through furin cleavage under consideration: 2018 onwards. The surprising failure to find better evidence for a natural spillover, and the lack of transparency from the Chinese scientists, is therefore best explained by positing a laboratory accident involving a live virus experiment as the cause of the Covid pandemic and attempts to cover it up.
This is a Bayesian conclusion, arguing that the total weight of the evidence supports a lab-leak prior. And it sure sounds conclusive, but I’m wondering why the paper was rejected (they don’t say what journal they submitted it to).
Further, a number of virologists I respect either adhere to the alternative wet-market theory or remain agnostic. When I asked a colleague some questions about this, he/she said this:
All the **data** (including new stuff) points to a natural origin. It might have been a leak, but all the evidence that has been obtained points in the direction of a spillover in the wet market. Not everyone who disagrees with the prevailing view of something is Galileo.
And then I asked “What about the furin cleaveage site?” This was something that Nobel laureate David Baltimore considered almost conclusive evidence for the lab-leak theory, but walked it back a bit:
The virologist David Baltimore commented that “these features make a powerful challenge to the idea of a natural origin for SARS2,” later clarifying that “you can’t distinguish between the two origins from just looking at the sequence” (Caltech Weekly 2021).
That article, in the Journal of Virology from 2023, concludes that the wet-market hypothesis is more likely, though is somewhat agnostic:
Scientific conclusions are based on likelihood given the scientific data, and conclusions can change as new data are obtained. Based on the scientific data collected in the last 3 years by virologists worldwide, hypotheses 1 and 2 are unlikely. Hypotheses 3 and 4 cannot be ruled out by existing evidence. Since hypotheses 1 and 2 support the lab leak theory and hypotheses 3 and 4 are consistent with a zoonotic origin, the lab leak- and zoonotic-origin explanations are not equally probable, and the available evidence favors the latter. Further insight into CoVs in animals at the animal-human interface requires additional surveillance of circulating virus sequences from animals. There is ample precedent for the seeding of pandemics and more geographically limited outbreaks from nonhuman species. Common-cold CoVs, SARS-CoV, Ebola virus, HIV, influenza A virus, mpox virus, and others all have zoonotic origins (31–33). SARS-CoV-2 is the ninth documented coronavirus to enter the human population. The best existing scientific evidence supports a direct zoonotic origin. As new evidence continues to emerge from scientific studies or other investigations, our understanding of the origin of SARS-CoV-2 will continue to evolve. Nevertheless, it is possible that its origin may never be known with certainty.
I’ll add that Ridley is, as he says below, an adherent to a less stringent form of climate change than many scientists. He calls himself a “climate-change” lukewarmer (see comment #2 below). Of course, he could be wrong about that yet right about covid. Nevertheless, I remain agnostic about how the virus got into our species. We may never know how this occurred, but investigating its origin is still worthwhile. In the article cited above, David Baltimore explains why:
Why is it important to know where the virus originated?
Well, I think we want to know the pathway of generating highly infectious new viruses that could cause pandemics because we want to protect ourselves against this happening again. If it happened by natural means, it means that we have to increase our surveillance of the natural environment. We have to try to find the hosts that provide an ability for the virus to change its sequence, to become more infectious. This would mean we need to keep surveillance on markets, on zoos, on places where viruses could jump from one species to another.
But if SARS-CoV-2 came about by an artificial means, it means we’ve got to put better defenses around laboratories. I’m not suggesting that it was deliberately released if it came from a laboratory, but we have to realize that whatever a laboratory does might get out of the laboratory and create havoc. It means that work of this sort should only go on in what are called biosafety level 4 laboratories.
UPDATE 1: Author van der Merwe wants me to add another point, which I will, despite his rude last line:
I would add one more point, concerning why this issue matters.
Even the possibility that the COVID-19 was the result of a lab leaks mandates that close attention is paid to preventing another research related pandemic in future. At present scientists largely self-regulate when it comes to deciding what gain of function experiments to do, even if such experiment could generate a pandemic organism. Many of the scientists arguing most strongly for natural origins have a clear conflict of interest as they oppose tightening of restrictions on these sorts of experiment.
Refusing to tighten regulations until a lab leak is proven, is equivalent to refusing to reduce CO2 emissions until it is proven beyond doubt that these greenhouse gas emissions will result in catastrophic effects.
It is ‘pandemic-risk denialism’.
Are you siding with these people?
UPDATE 2. Ridley also wrote me, as expected. and wants me to change my article. I have modified my comment on climate change to reflect his claim that he is a “lukewarmer.” But his last point I already dealt with in my comment from Baltimore in the original post.
1. The (anonymous) virologist colleague that you consulted has misled you in a rather shocking way. He told you that furin cleavage sites are “common in closely related [corona]viruses.” This is simply untrue: the only coronaviruses that count as “closely related” by a reasonable definition of that term are other sarbecoviruses. None of the over 800 sarbecoviruses yet found have a FCS. Not one. Other less closely related coronaviruses do have one, as we state in our article, where we say “MERS has one”. But to call other coronaviruses, such a as merbecoviruses “closely related” to SARS-CoV-2 is like calling birds or reptiles “closely related” to a mammal. You would not argue that finding the first mammal to grow feathers was unsurprising because “closely related tetrapods have feathers”. And if it was found near a lab that had engineered feather genes into other animals, I would be surprised if you did not find it suspicious. Note that an FCS is selected against in bats, where sarbecoviruses cause enteric infections, as we state.
2. Your colleague says “all” the data points towards the market. That’s just an empty slogan: what data? Not a single infected animal has been found, the very minimum expected in every other zoonotic outbreak. We set out the data on both sides of the argument in our paper and lots point to the lab. None of the data is proof but that does not make it “un-data”.
3. You wonder what the editor said to justify rejection. He said that there is no evidence of gain of function research at the WIV. That’s simply laughable. There’s tons of published evidence of exactly that, proudly published in journals by the WIV scientists.
4. You call me a climate denialist. This is false, lazy (easy to check) and defamatory. I have covered climate change in print for more than 40 years. I have repeatedly stated exactly why I think climate change is real and man-made. I just don’t yet see good arguments for it being net very dangerous. That’s not “denialism”. I am a lukewarmer.
5. Several commenters argue that it does not really matter how it started. This is gobsmacking. Knowing how this pandemic began is vital for preventing the next one: next to no effort is being made to enhance lab safety currently, or to prevent terrorists gaining access to virology expertise. In addition, a lab leak could explain why the virus was so highly infectious from the start of the outbreak. I have yet to hear anybody argue after an airliner crashes that it does not matter what caused it.
I suggest your readers read the whole paper.
h/t: Christopher for the Torygraph archive.
The fact that the virus appeared right next door to a viral research lab gives a likelihood ratio heavily favouring lab-leak, and you’re going to need a lot of evidence in favour of zoonotic origin to overcome that.
If Covid had first been seen next door to Porton Down then we would all have assumed it came from there, and there would have been a massive public inquiry.
Meh. The virus was also found “right next door” to a place that had live, infected animals that people were handling and eating.
Yes but there are vastly more wet markets than there are viral research labs, so the LR for that piece of evidence heavily favours lab leak.
Viral labs are not located randomly – some are located where potentially dangerous viruses exist in abundance locally and can be collected easily, which is why the Wuhan lab is located where it is.
The statement:
“The fact that the virus appeared right next door to a viral research lab gives a likelihood ratio heavily favouring lab-leak” is doubtful.
Does any spatial correlation affect the LR, though, as the competing hypotheses are already conditioning on lab source and wet market source?
Evidence: In Wuhan
H_p: from a Lab
H_d: from a wet market
Then, admittedly simplistically:
LR = P(E|H_p) / P(E/H_d) = (1/#labs) / (1/#wet markets)
= #wet markets / #labs = large value => favouring H_p
In other words, assuming it’s from a lab, what’s the chance it’s this one, versus assuming it’s from a wet market, what’s the chance it’s this one?
OK I’ll shut up now to avoid oversharing, I’ll just note that the above is an example of a standard approach to weighing evidence that you must teach if you want your forensic science course to be accredited.
Of course you could argue that the hypotheses should be “this lab” and “this wet market”, but then you are conditioning on the outbreak being in Wuhan, rather than treating Wuhan as evidence.
Actually, spatial correlation will upweight the chance of it being froma wet market that is near a lab, so the denominator in the LR will be a bit larger and hence the LR a bit lower than thaeabive naive version.
I suggest that anyone interested in the topic go to:
https://www.microbe.tv/twiv/?s=covid+origin
and start listening to the discussions by virology experts, including those doing research on COVID19 origins. You will discover that the first bullet point by Matt Ridley:
“Attempts to find evidence for a wet-market leak have been unsuccessful. The cases found around the wet market could simply reflect sampling bias, as the Chinese concentrated on looking for infected people in that area.”
is not accurate.
The bats that harbour SARS-CoV-2 like viruses live over 1000 km away from Wuhan. That is where the closest known relatives of SARS-CoV-2 are from, including on virus brought to Wuhan by scientists. The only known way they could move between the two is scientists, who were trafficking viruses from these bats to Wuhan for over 8 years before the pandemic.
“The virus was also found “right next door” to a place that had live, infected animals that people were handling and eating.”
The investigations have not found this (based on my reading up until perhaps a year ago; I do need to read more). All virus traces in the “wet market” were determined to be of human origin (virus shed by humans). No infected animals were found.
Not true. No infected animals were found at the Wuhan market.
But since there are no examples of a lab-generated virus escaping and causing disease in the history of the world, that scenario is overcome by the natural origin one which accounts for every outbreak to date.
That’s the fun of Bayesian theory.
🙂
Marberg virus escaped from a lab in Marberg Germany. I understand there have been several cases like that. AI says:
“Several viruses have been implicated in laboratory-related incidents, including escapes and laboratory-acquired infections (LAIs). Some notable examples include Bacillus anthracis, SARS-CoV (the virus that causes SARS), poliovirus, Brucella species, foot and mouth disease virus, variola virus (the virus that causes smallpox), Burkholderia pseudomallei, and influenza virus H5N1.”
Not saying this means covid escaped from the Wuhan lab, just pointing out that viral escape from labs is not unprecedented.
This:
“But since there are no examples of a lab-generated virus escaping and causing disease in the history of the world, that scenario is overcome by the natural origin one which accounts for every outbreak to date.”
Is tantamount to saying no infectious agent has ever escaped a laboratory and infected anyone outside the lab. That is pretty absurd on its face. You are claiming a perfect record for labs using infectious agents. That is incorrect.
Just one for instance: https://en.wikipedia.org/wiki/Marburg_virus#Human_disease
In addition, has GOF testing been performed on SARS-family viruses, particularly WRT to infectiousness in humans, previously in the history of the world? I think the answer to that is no (and I hope it won’t be performed in the future). New experiments, new results. And: BSL-2 (FFS).
The research lab is a 40 minute drive from the market, on the other side of the Yangtze River. I would not call that “right next door”
Check out the map here:
https://www.bbc.com/news/world-asia-china-57268111
The first cases identified were not “right next door” to a SARS coronavirus research lab. They were instead found in the same city as a research lab, quite a distance away from the lab. Moreover, a densely populated urban environment helps SARS-CoV-2 transmit. That makes it more likely that initial cases found would be in a city. And many cities in China have labs doing research on SARS-like coronaviruses. So it’s unsurprising that the first cases found would be in a same city with a SARS research lab, even if the outbreak was zoonotic. In fact, Ridley’s original lab leak claims focused on a site that was not primarily a SARS research lab.
-“The earliest tweets I can find [from Ridley] are actually not blaming the Wuhan institute of Virology, but rather, the Wuhan CDC.”
https://threadreaderapp.com/thread/1925696479747867105.html
-“13/16 Turns out most major cities (Hong Kong too) have sites that could & would be blamed, after the fact, for a pandemic. There’s even a lab in Beijing that collected bat virus samples from the ‘Mojiang mine’ and worked with @PeterDaszak and EcoHealth Alliance. A few examples:”
https://threadreaderapp.com/thread/1633572396639862785.html
-“The origins of SARS-CoV-2: A critical review
[…]
The suspicion that SARS-CoV-2 might have a laboratory origin stems from the coincidence that it was first detected in a city that houses a major virological laboratory that studies coronaviruses. Wuhan is the largest city in central China with multiple animal markets and is a major hub for travel and commerce, well connected to other areas both within China and internationally. The link to Wuhan therefore more likely reflects the fact that pathogens often require heavily populated areas to become established”
-Trevor Bedford: “But keep in mind, that there are other labs in China (if this had originated in Beijing, people would be accusing China CDC instead of the Wuhan Institute of Virology)”
-“Confirmation of the centrality of the Huanan market among early COVID-19 cases”
-“Dissecting the early COVID-19 cases in Wuhan”
– “The Huanan Seafood Wholesale Market in Wuhan was the early epicenter of the COVID-19 pandemic”
-“No evidence of systematic proximity ascertainment bias in early COVID-19 cases in Wuhan Reply to Weissman (2024)”
So you had a viral research lab and a wet market in proximity to one another. It is plausible that the virus might have originated in either place, so saying “it had to be the lab”, in lieu of any other evidence, seems to be fallacious reasoning to me.
But let me be clear: I have no idea which it was, and finding the truth is always OK with me.
My bias (well, one…): the way that humans raise and grow animals for food creates and environment for virus to mutate and become transmissible to humans. I think there are many people who don’t want to face that fact, and the lab leak theory is their pet.
OTOH, gain of function research is scary. Should we really be doing that anywhere?
China shut down the wet market and cleaned it out before anyone could sample animals from there. That is why the wet market evidence is not as strong as it could be. The sampling was of areas where animals were sold, but the animals were already gone. I think China knew it likely started in the wet market and that is why they shut it down.
The animals were gone before anyone showed up to sample the market and before the market got shut down. Word gets around fast these days and people in an illegal industry knew to shut it down ASAP. Same thing happens regularly to stay ahead of inspections.
Yes, and still there were still samples of early Covid from that one corner in the market, in hair and fecal samples. I think it was where Raccoon Dogs were kept.
I have not read of this data being presented. From my reading, all the virus at the “wet market” was determined to be shed by humans.
I have not read that data being presented. How would they know it was shed by humans? Now I have read my claim from different places, but here is one: https://www.nature.com/articles/d41586-024-03026-9 (probably pay-walled, but …)…
“The genomic data used in the Cell, Nature and other analyses were collected by researchers at the Chinese Center for Disease Control and Prevention (China CDC) shortly after the market was shut down on 1 January 2020. Over several weeks, China CDC staff visited the market many times to swab stalls, rubbish bins, toilets, sewage, stray animals and abandoned frozen animal products. The samples contained lots of DNA and RNA from multiple sources that researchers had to sequence and sift through.
“It’s one of the most important data sets on the early pandemic and on the origin of SARS-CoV-2,” says Florence Débarre, an evolutionary biologist at the French national research agency CNRS, and co-author of the Cell analysis.
When researchers at the China CDC published their analysis in Nature last April, they reported samples that contained SARS-CoV-2 and came from wild animals in the market, most notably raccoon dogs (Nyctereutes procyonoides), which are susceptible to SARS-CoV-2 and can spread the virus to other animals. But the team noted that there was no way to establish that the animals were infected with SARS-CoV-2. Even if they were infected, they could have caught the infection from a person who brought the virus to the market, which leaves open the possibility that the market was not the site of the pandemic’s emergence.”
That last bit could be about your claim? So either there is a NEW FINDING (possible), or what you said is from a muddled source (maybe also possible?)
More: “The co-location of viral and animal genetic material is “strongly suggestive” that the animals were infected, says Gigi Gronvall, a biosecurity specialist at Johns Hopkins University in Baltimore, Maryland. “I was quite amazed by how many animals were there,” she says.”
And: “The authors of the Cell study also argue that the viral diversity present in the market suggests it was the site of the pandemic’s emergence. In particular, they say the presence of two SARS-CoV-2 lineages — known as A and B — circulating in the market suggests that the virus jumped twice from animals to people. The researchers conclude that, although it is possible that infected humans brought the virus to the market on two separate occasions, that is a much less likely scenario than the virus jumping twice from animals, especially since their analysis suggests that very few people would have been infected at that point and it is unlikely that one person seeded both lineages. “It really just fits this ongoing infection in animal populations that spilled over multiple times to people,” says Gronvall.”
I don’t think we know who shed the virus. We do know that all have the virus that was found was strains adapted to human transmission. There was no virus that was particularly adapted to Racoon Dog (for example) transmission.
I assign my engineer’s comment and opinion proxy to Hempenstein – a life sciences subject matter expert.
IF there is anything new in the jargon. Thus far I am a when I hear hooves I think horses, not zebras kind of guy….and see TWiV and Paul Offit as trusted sources.
The TWiV guys and gals have covered the lab leak vs zootic spillover hypothesis. They have an excellent YouTube presence. Also at microbe.tv. Data and evidence driven.
Just completed the article. The furin cleavage site seems to be strong evidence in favor of the lab leak hypothesis, and the arguments for the other hypotheses seem to be weak. Is the furin cleavage site a smoking gun? Well, maybe or maybe not. The fact that it is not normally present in SARS viruses and that it seemingly appears in SARS-CoV-2 just in time to start a pandemic tilts my current thinking towards the lab leak origin but, not being an expert, I’m not sure I should be keeping score.
Try this:
https://pubmed.ncbi.nlm.nih.gov/33340798/
See, it’s like all over the g.d. place all in this virus family! And viruses pick up genetic material all the time. There is no excuse for dismissing this point!
So, it could be of natural origin. But it also could not be of natural origin. The fact that it occurs naturally (the scientists who may have inserted it wouldn’t have tried that if this weren’t the case(!)) does not prove that it was of natural origin.
I’ve never said that it’s 100% certain to have been a lab leak. The inverse (100% certainty of zoonotic spillover origin) seems to be the position of most of the critics of Ridley and Chan.
I fully agree that the default should be zoonotic origin, based on previous known events of virus spillover.
However, that does not preclude an accidental lab leak.
As I’ve noted elsewhere in this thread I need to read the purported trove of new evidence for the zoonotic origin hypothesis.
I can agree. Furin cleavage is not demonstrative either way so both scenarios are possible. But lab leak leaners don’t seem to admit that.
It’s 12 nucleotides. It’s the smallest God of the gaps in history and no self respecting intelligent design theorist would propose such a thing.
Coronaviruses acquire whole open reading frames and no one can tell you where they come from. Acquiring 12 nucleotides that encode 4 amino acids in a disordered loop that’s the site of strong selective pressure is boring.
You don’t need to ask me. You can ask… Matt Ridley. He wrote in his book with Alina Chan that the furin cleavage site didn’t tell you if the virus was natural or not, writing, “finding a natural origin of the furin cleavage site will not clear up the question of whether the virus first jumped into people in the wild or in the course of research activities.”
Ridley and Chan’s dilemma is that the furin cleavage site is the only oddity they cherry picked that’s still not been found in nature. So now they need to pretend that it’s a smoking gun.
SARS-CoV-2 in humans acquires similar inserts all the time — almost never can anyone tell you with any confidence how it happened.
Lastly, Ridley bookends his article by first arguing that SARS-CoV-2 is suspiciously well adapted — furin cleavage site and so on — and then arguing that SARS-CoV-2 is suspiciously maladapted — because of an early mutation that increased fitness. It’s a hilarious Goldilocks principle of detecting the “just right” level of adaptation that characterizes allegedly synthetic organisms.
The whole thing is a total joke. I don’t begrudge people at all for falling for it given how little skepticism there is in the media for lab leak “sleuths” who are just self-promotional conspiracy theorists.
From my reading of Ridley and Chan, this isn’t their point(s) (this is from my reading of Viral, Ed. 1.
They state that, unlike SARS-1 and MERS, the virus did not exhibit rapid and significant evolution to adapt to the human host, suggesting that it was pre-adapted.
They suggest that the furin site was added as part of GOF testing at WIV, and was tested there to determine that the change in fact increased its infectiousness in cultured human cells.
I’m going to look at some of the links presented in this thread. There’s much talk of new evidence for the zoonotic spillover hypothesis. I need to read that.
I’m glad you are looking at things more carefully that I am. I wonder if the finding that the virus was not evolved to be adapted to human hosts so well refers to the codons that the virus was using. The virus tended to use codons that humans don’t prefer (although they still work in humans), so it would not replicate as efficiently in humans as a virus would if it evolved for a time in us.
Well, that furin cleavage site is still sufficient for it to jump to humans, and it could still replicate in us.
Re the furin cleavage site appearing “just in time to start a pandemic”: the zoonotic hypothesis would have the (natural) gain of that site (likely from horizontal transfer from some other virus) as being a critical event in the triggering of the pandemic.
There’s a detailed critique of Ridley’s article, including with respect to its claims on the furin cleavage site (FCS). The evolution of the FCS has already been discussed in the literature + by other commentators, via mechanisms such as copy-choice error. Its existence is not indicative of a lab leak:
-“The origins of SARS-CoV-2: A critical review
[…]
simple evolutionary mechanisms can readily explain the evolution of an out-of-frame insertion of a furin cleavage site in SARS-CoV-2”
-“The SARS-CoV-2 furin cleavage site: natural selection or smoking gun?”
-“SARS-CoV-2 furin cleavage site was not engineered”
-“The evidence remains clear: SARS-CoV-2 emerged via the wildlife trade”
-“The emergence and evolution of SARS-CoV-2”
-“There is still no evidence of SARS‐CoV‐2 laboratory origin: Response to Segreto and Deigin (10.1002/bies.202100137)”
https://virological.org/t/tackling-rumors-of-a-suspicious-origin-of-ncov2019/384/1
https://threadreaderapp.com/thread/1925696479747867105.html
As I understand it the paper was rejected because it speculated that human-specific furin cleavage site insertion at the S1/S2 boundary had taken place at Wuhan and there is currently no concrete evidence that this is the case (but nothing wrong with speculation if it is clearly stated as such). But of course the rejected DEFUSE proposal had planned to do exactly this at Wuhan. And then such a virus appears in Wuhan. This is why the Bayesian prior at the Wuhan lab is so much greater than at the wet market.
How can un-evidenced speculation (that Wuhan lab researchers were doing research they weren’t supposed to be doing) strongly tip the scales in a Bayesian analysis? What kind of quantity could be added with that?
Consider the situation before the outbreak, and the Bayesian priors on
all the possible places where the next new virus to which humans would
be vulnerable might appear. These could be wet markets, or houses of
people who have been to distant caves with bats, or whatever else you
can think of. There might be thousands of possible places each with
Bayesian prior of say p=0.0001. One of these is in Wuhan – the wet
market. And now we need to provide a prior for the Wuhan lab, a place
where they wanted to do gain of function experiments, including ones
tailored for infection of humans, but they hadn’t got funding. (As I
understand it this was the only lab in China where this work might
have been done.) We don’t know for sure that they were doing the
DEFUSE experiments so we have to make a guess if it was
likely. It is not uncommon in Bayesian analysis for the prior to involve a fair degree
of subjectivity, and in this case it is inevitable that it involves guesswork. But the key point is that it’s conceivable that they were doing the work and so the prior probability is not zero. Let’s
say there is a 1% chance the next outbreak will due to leakage from
the Wuhan lab p(lab)=0.01, given that they wanted to do the work, and
that they discussed doing experiments at Biosafety level 2. The thing is in
a Bayesian analysis you have to subjectively choose a prior. You can’t
say ‘we can’t be certain, this is un-evidenced speculation, so there
is no information’. Now to be concrete we will say that p(lab)=0.01 and there are 9900 other possible sites with p=0.0001.
So, a priori the chances of the next great outbreak happening at the Wuhan lab is only 1%. It’s much more likely to occur at, say, a wet market somewhere. At this point, my guess is that the situation summarised in the previous sentence largely conforms with what most people intuitively believe.
But now the outbreak happens and it’s at Wuhan, and there are only two possible sites. One is the wet market with prior p=0.0001, and one is the lab with prior
p(lab)=0.01. The posterior probabilities become p(market)=0.01 and p(lab)=0.99.
[The posterior probabilities in all the other sites, outside Wuhan became zero.]
[This is not very different to TJR’s argument above]
Ridley kicks off the introduction to his manuscript with a lie. He says he and Alina Chan were on the fence when they wrote Viral, until “in the autumn of 2021 more startling evidence emerged to support the lab leak. I now think that is by far the most likely explanation.”
The “startling evidence” is the DEFUSE proposal. This proposal is described at length in the first edition of Viral. It wasn’t new evidence; he’s just a liar.
While working on editing the second edition, Alina Chan asked for advice from someone who wrote a mixed review of the first edition: “Would proposing that a lab #OriginOfCovid is the most likely origin make the book less scientific but more impactful?”
And that is exactly what she and Matt Ridley did.
The fact is that for a year starting in Spring 2021 a ton of evidence came in that disproved all of the mainstream lab leak theories that were to some degree specific enough to be falsifiable. Lab leak theorists just pretended that never happened and asserted that they were more confident than ever. Evidence has continued to come in since then and has been perfectly consistent with origin in the wildlife trade without any surprises that point anywhere else, lab or otherwise.
This comment strikes me as quite unfair to the authors. As someone who trained in the field (though I no longer work in it), I closely followed Matt Ridley’s public statements during that period—particularly on Twitter—and my memory of his evolving position is consistent with what he described at the top of the article.
The claim that DEFUSE was “described at length in the first edition” is, in my view, misleading. The DEFUSE proposal was leaked only in mid-September 2021, just weeks before the book’s early November release. By that point, the manuscript had already been submitted for publication. DEFUSE is discussed only at the end of the epilogue, across three short paragraphs — a section that was clearly added at the last minute (standard practice, I would assume, for incorporating breaking developments).
I feel that is further supported by the differences between the UK and US editions of the book (I have both). The UK version, published slightly earlier, is visibly more tentative: it actually doesn’t even mention the word ‘DEFUSE’ and refers to it as a “purported” proposal, reflecting early uncertainty about its authenticity. The US edition, printed a couple weeks later, is more confident and attempts to analyze the meaning in a broader context — indicating that the authors were still actively processing and incorporating the implications of the DEFUSE leak as the book went to press.
I also feel that the differences between the versions support the idea (which I also feel is consistent with Ridley’s public commentary at the time) that the appearance of the DEFUSE document had a significant impact on his thinking, as it provided concrete evidence that certain types of high-risk research had in fact been proposed by major players. I am not saying that he went from 50/50 to strongly on one side. Throughout the course of researching and writing the book he had shifted towards feeling that the lab leak was more likely than not (again I would assume quite common for people writing a book on breaking story). This was significantly enhanced with the DEFUSE document that released as he was finishing the final draft of his epilogue.
There has been no evidence that disproves the lab leak theories — indeed the evidence in favor has just become stronger and stronger. The structure of the COVID virus remains consistent with what would expect to see if the WIV proceeded along the research path described in the DEFUSE grants.
And, on the other side, there has been no evidence that the the virus was spread through the wildlife trade. And, there are lots of surprises that continue to argue against the wildlife trade (like the fact that there is no evidence of spread during transit). I am not sure what you are referring to when you say that there has been lots of evidence.
Rejection of a paper is usually triggered by multiple inferential flaws. I don’t know this subject at all, but it is striking to see these inconsistent arguments in the paper, as summarized in this post:
“The cases found around the wet market could simply reflect sampling bias, as the Chinese concentrated on looking for infected people in that area.”
“If a person got infected with a bat virus from Yunnan (one theory), that person would have infected others on his/her journey to Wuhan, but there is no such trail of infection.”
If the first argument is true then the second argument is false.
This example suggests to me that the paper is employing motivated reasoning and cherry-picking to support a predetermined outcome. The lab leak theory does seem to me to be the most likely explanation, based on the totality of evidence, but it is a tricky call. I suspect this paper is not dispassionately analyzing the evidence, and this may have been the reason for its rejection.
While I think that this line of thinking shows good instincts, I do not feel that the two statements are necessarily inconsistent. In the case of the first one, it is a caution against over-reading a pattern: just because many cases were found near the market doesn’t prove the outbreak started there. Whereas the second one is an attempt to assess an alternate theory: if the virus came from Yunnan via a traveler, then some evidence of infections along the route might be expected. However none has been found (or at least none has been disclosed). I view that as more of a Bayesian weighing of alternatives, where each line of evidence has limits.
I’m not sure I understand. I was thinking that if one argues that lack of sampling explains the absence of evidence in the first case, well, it also would explain the lack of evidence in the second case. Yet the author holds up the lack of cases along the route from Yunnan to Wuhan as a reason to reject that route of infection.
He argues (perhaps correctly) against over-reading a pattern in the first case, but then himself over-reads a pattern in the second case, probably because it fits his narrative.
Well, I was characterizing what I got from the paper, and maybe people who are invested in this should actually read the archived version (I haven’t, so I don’t know if it’s identical to the one in the Telegraph).
Those two statements are not at all contradictory. The first refers solely to the public health response in Wuhan — where officials focused on the market. The second refers to the fact that the virus appeared in Wuhan but not in any of the cities between Yunnan and Wuhan (which would not have been subject to bias by Wuhan officials). That is clear from the statements and there is absolutely nothing contradictory about them — I can’t imagine why you think there is.
If sampling bias can explain the first observation, why does it not explain the second observation? It’s not that the observations are contradictory. The problem is that the same counter-argument could explain both of them, but the author chooses to use it when doing so supports his thesis, and avoids it when doing so hurts his thesis.
That’s my last word on this, fide Da Rulz…
Anyone else puzzled by the passion that this issue generates, given the lack of dispositive evidence? Of course, maybe I shouldn’t be. We have people today who are enthused about teen girls slicing off their breasts and men punching women in their faces for accolades. What evolutionary history connects these supposedly disparate phenomena?
“What evolutionary history connects these supposedly disparate phenomena?”
Nothing at all. The passion is from politics, nothing more. Nothing adaptive about that. Remember, when you mix science with politics all you get is politics, and like religion, politics poisons everything.
And you don’t believe the tribalism and other behavioral maladies plaguing our politics have an evolutionary history?
So…your question was rhetorical?
But why is this political? I don’t regard the lab leak vs. wet market hypotheses as political in the least, but evidently others do – with most of the passion coming from those who oppose the lab leak hypothesis based on what I’ve seen. Why is it so upsetting to some people to even consider a lab leak origin for covid? It isn’t as though nothing like that has happened before (see my earlier post above). And I’m not referring to those who suggest the lab leak was deliberate, which I find highly implausible.
I feel that for many people, this issue has become purely political. But for me, it’s not political at all. It was a sad and painful situation. One where I had to come to terms with the fact that many of the people I once held in the highest regard acted in ways I believe were contrary to the public good and damaging to public trust in science. That sense of disappointment remains, regardless of where the evidence on the origins ultimately points (if we ever have enough evidence for it to strongly point in any direction). I’m still somewhat agnostic on that question.
It’s a good question. Why should we get political or at least emotive about this debate — one that probably will never be resolved?
I am a pro-science guy with a fair amount of experience in the drunkards’ walk of how science really works, and how the process sometimes gets to a robust consensus eventually.
Pretty much all of the evidence put forth for the lab leak was just soooo much innuendo and leaps to conclusions. The proximity of the lab. The lab wanted to do gain of function research. Dr. Fauci, who had a research interest in the lab, was (gasp) mean to people promoting the unjustified innuendos at a time when humans needed to goddam trust the scientists while they worked like hell to get vaccines out. And f.f.s stop taking horse de-wormer!
The research was messy. But the disinformation was worse.
Enjoying this thread. I am glad Jerry says he is on the fence about this issue (I wasn’t so sure before)!
I count myself as one who favors zoonotic origin, given known facts which have been described here many times before. But also I don’t think zoonotic origin is proven without doubt nor that lab leak is disproven without doubt. I don’t know of any case where a significant virus’ origin has been pin-pointed to any degree like we seem to demand of this one. Yet I think ALL other significant viruses are zoonotic, are they not?
Finally, lab leak proponents too often let slip their biases with slanted and nakedly truncated background about virus biology, proximity of the lab to the wet market, Chinese secrecy, and so on. That does not disprove the lab leak, but it does reveal the lab leak proponents biases.
The biases come from the data. There are wet markets all over China but only one virology lab doing gain of function research on bat coronaviruses and that was in Wuhan. It does on the surface look like Occam’s Razor applies. That does not mean it does apply in this case – it could just be an amazing coincidence.
I am tired of posting this: the most convincing argument that I have seen from either side, and it comes down strongly on the side of a lab leak:
https://www.scribd.com/document/492682888/SQuay-Bayesian-Analysis-of-SARS-CoV-2-FINAL-v-2#
It is a very thorough Bayesian analysis, starting with a conservative prior of 0.012 and concluding with a probability in favour of the lab leak of 0.998. The probability estimates that lead to this seen to me to be conservative.
I would really like to see a serious refutation, if it is possible. Please can some wet market enthusiast oblige?
Wish I could, since I want to see something other than innuendo from the lab leak side. But is there a way to see this article without being pushed toward a paid subscription?
193 pages???
I managed to download a copy a couple of years ago. I seem to remember that it was not straightforward, and I cannot remember how I did it, unfortunately.
Yes, it is long, but that is the point: it is very thorough, much more so than any other so-called analysis that I have seen..
It seemed to me that the way to do that is thru a free trial. I don’t trust those things, especially that I’d forget to cancel or something.
Me again. So I can get to the early big point that the earliest patients found to exhibit pneumonia-like symptoms were found to have Sars-Cov-2 sequences but also sequences for an “adenovirus pShuttle vector”. This is described as a smoking gun for genetic manipulation. This actually seems very important toward lab leak, so I wonder why lab leaks proponents aren’t hollering about it. This is the first I’ve read about it, at least.
Now I know jack all about this stuff. A Google search for Covid infections and adenovirus turns up lots of papers about how people readily get co-infected with both kinds of viruses. But that might not mean anything here.
A narrower search turns up a lot of articles about how efforts to develop vaccines during the pandemic included the use of vaccines based on adenoviral vectors.
That’s all I got! I’d think there would be a lot of news (and a lot of Google ‘hits’) about this coincidence with these early patients. But several searches turn up nothing else.
All labs have leaks.
Labs working at BSL-2 are far more likely to leak that labs working at BSL-4. I find in interesting that none of the critics of Ridley and Chan in this thread are addressing this issue.
“Why is it important to know where the virus originated?”
Both scenarios of “why” to me are like battling climate change. Unless all countries are dedicated in the endeavor of stopping another pandemic, it’s pretty much a lost cause. I don’t know if China has changed any of it’s protocols/security in this area, and if they did succeed in manipulating the virus, I doubt they stopped the research. I’m pretty sure wet markets are still active as well. America’s current government couldn’t give a shit.
I doubt there will ever be consensus on the origin story of Covid. I’m sure something similar will happen again since the root cause: wet markets, world ecological degradation and/or human error won’t be addressed in any serious manner. And there’s no stopping evolution.
Cue the theme song of 12 Monkeys …
It still seems to me that the major evidence that convinces most people of the lab leak theory is that Wuhan Institute of Virology and Wuhan Wet Market start with the same name.
Try a quick thought experiment and switch Wuhan Institute of Virology with Shanghai Institute of Virology but keep the wet market in Wuhan.
For the animal origin people this makes zero difference to their conclusion. All the evidence that convinces them (I’ll post three) is still there
1: All initial cases cluster around the Wuhan market and not a different focal point
2: Sequencing of samples from the market finds high levels of the virus in animal cages that housed animals now known to be susceptible -racoon dogs (DNA from the animals was also detected in the cages with high viral counts)
3: Sequence from the two earliest lineages was found in both early human cases from around the market and in the cages (indicating that the virus was present and evolving in the market, presumably within the animals before at minimum two jumps to humans).
Now this does not discount the lab leak theory. We could have one infected researcher travel from Shanghai to the Wuhan market and sneeze on a racoon dog a month before the outbreak (giving enough time for the two lineages to evolve), or alternatively, two researchers from Shanghai, infected with different strains, visit Wuhan and sneeze on the same poor racoon dog.
Ridley and Chan concentrate on the first point (possible sample bias around the market) but avoid the molecular smoking gun of the two lineages within the market cases and animal cages.
I the lab was in Shanghai rather than Wuhan, would that make you more or less confident in the lab leak theory? It still started spreading in the Wuhan market but there is as much evidence that it started in a Shanghai research lab as there is that it started in the Wuhan Institute of Virology(i.e. zero).
Regarding lab leak, I can only say that it must have been seen a major possibility at the beginning (lab leaks have happened in the past – although in the era of genomics they are now generally traceable to the point of origin). The three points of evidence I listed make that lab leak theory much weaker and a natural origin far stronger (basically a rerun of the origin of the original SARS).
SARS-Cov2 was an unknown virus at the start of the outbreak- definitely not a good candidate for gain of function experiments. There is no evidence that it is engineered (the furin cleavage site is common in coronaviruses).
One possibility is that they were collecting and storing different bat viruses and there was a leak (aerosolization of a culture and transfer to a worker). That seems far more likely than a gain of function product or deliberately engineered virus (and is consistent with what we know they were doing at the WIS.)
However, like the Shanghai thought experiment, we still have to address the three points I listed.
I think you haven’t seriously read any proponents of the lab leak theory if you think the similarity in names is meaningful to anyone. Also, if you had seriously read the proponents, you would know that there are very strong counter-arguments to the three points you make. And, it appears that you are entirely unaware of the DEFUSE grant that shows that the WIV (not some random lab in Shanghai) was planning on conducting research that could easily lead to the creation of SARS-Cov-2. Rather than being quite so pompous, you should try reading people with different views.
Well, since both names are ’Wuhan’ obviously means they are both in the same city, I would suggest this IS convincing for many (just read the first sentence in reply 1 to this story).
Let’s take a different thought experiment route and accept the premise of the lab leak theory.
The WIS either engineered SARS-Cov2 from a previously unpublished bat virus (we still do not have the sequence of the un-engineered virus) and accidentally released it, or they were collecting and storing different viruses and accidentally released one.
The molecular evidence we have (viral lineage trees from environmental sampling within the market) still points to a few cages in the market that had housed raccoon dogs as being the likely initial focal point of spread.
So, the thought experiment is how do we reconcile these two points (that there was a lab leak, and that the spread began within the market).
The problem I and many scientists have with this is that it requires two leaps of faith, that there was a lab leak, and that the spread began in a manner that would normally make people assume that it was a zoonotic spread.
Either this was a deliberate engineering of the evidence after the fact to make people think it was a zoonotic spread (putting viral samples of both early lineages into the cages, in addition to genomic DNA from raccoon dogs, or that it was some crazy coincidence (like the WIS testing new bat viruses on raccoon dogs in their lab and selling the animals in the market immediately afterwards.)
Each of these allows for the lab leak but also allows us to keep the known molecular genetic evidence from the market.
Is there a simpler explanation for the lab leak that allows us to keep the genetic evidence?
Occum’s Razor…Natural transmission happened many times over history. Simplest explanation is it happened again with Covid-19 virus.
Is Occam’s Razor always correct? Or is it just the beginning of an inquiry not the end.
“Related CoVs have FCSs” – this has been repeated ad nauseum by vocal “experts” during the past five years. I was unsure what to make of it until I read Ralph Baric’s transcribed interview (in 2024) with the Congressional Subcommittee investigating the pandemic.
As one of the world’s leading CoV researchers, Baric not only knew that other beta CoVs have FCSs, he had, in fact, shown that it played an essential role in MERS’ infectivity.
On page 193, Baric explains his motivation behind the DEFUSE proposal in this remarkable statement:
“We were fundamentally interested in why sarbecoviruses didn’t have a furin cleavage site”.
Baric’s interview: https://oversight.house.gov/wp-content/uploads/2024/04/Baric-TI-Transcript.pdf
I somehow missed this yesterday.
The most recent paper I’m aware of on this, with 20authors, came out in Cell last month and was discussed in TWiV 1217 (link below in a comment). In short, close relatives of SARS-CoV-2 were found in SE Asia, in horseshoe bats.
This is a dense paper and some of the terminology is outside my comfort zone, but what they argue is that natural mutational rates could get you from those SARS-CoV-2 relatives that they found over 1000km the Wuhan market, to the market, in ~10yrs, but the dispersal rate of horseshoe bats is too slow for that to have happened, and so interactions with intermediate hosts through the wildlife trade led to zoonotic spillover.
Link to TWiV 1217. Start @ ~25:00.