@George
Here's one other species that took that trip; certainly one of many. Not only from Yunnan to Wuhan, but from Yunnan to the same corner of the same market where the earliest COVID-19 cases and all genetic diversity of SARS-CoV-2 was found. The only place known to meet that description in the whole world other than, shortly after, healthcare facilities in Wuhan [1].

I'm surprised that Ridley hasn't read the WHO report given that he wrote a whole book on the subject. Good rhetoric to dupe gullible people on X, though. I'll hand it to Ridley for being better than the average lab leak advocate when it comes to speaking persuasively despite either knowing little or lying about what he knows.

[1] It's often said that markets are the natural place to first find an outbreak. That's only true if the spillover is in a market. Otherwise, the most likely place if the spillover is missed is in healthcare settings. This was the case in Wuhan with one exception: Huanan market. If someone shows up with the rootclaim argument about Thailand and Singapore markets well into 2020, get real. That's when the rest of the high density economy is shut down and/or heavily social distancing. Food is one of the few industries that couldn't do much of that, and 2020 outbreaks were unsurprisingly concentrated in food processing and distribution.

https://bmcgenomdata.biomedcentral.com/articles/10.1186/s12863-024-01290-2

Pre-pandemic artificial MERS analog of polyfunctional SARS-CoV-2 S1/S2 furin cleavage site domain is unique among spike proteins of genus Betacoronavirus

https://jimhaslam.substack.com/p/baric-referenced-a-2017-molecular

Baric referenced a 2017 molecular blueprint for COVID-19

https://youtu.be/EmWt00ak49k?si=sMb8ZWY7Pmp7E2a0

In German. I will create an English translation soon.

Lisewski reveals explosive details about the origin of SARS-CoV-2: Furin cleavage site, PAT7 sequence & genetic anomalies that raise questions. Here is the paper:

https://bmcgenomdata.biomedcentr...

I’m sure this must be wrong…🙄

@George so this is a precise molecular blueprint and Bruttel et al found a totally different precise molecular fingerprint.

In reality both are examples of the Bible Code applied to a genome sequence.

@zcoli if someone felt like wasting a lot of time, they could take a look at some other reasonably rare and similarly divergent virus sampled in a potential intermediate host, spend years cherry picking rarities and coming up with stories to explain how God put them there, and see that this isn’t unique to SARS2.

A reminder that there’s a set of questions on this topic that will actually resolve. I’m not betting on my own markets, but, in my opinion, it’s a good money making opportunity to vote yes.

My reasoning has nothing to do with expecting people to make a coherent case for lab leak and everything to do with the last month of American academia proving it will do whatever the President asks to keep their heads down and avoid becoming a target for spending cuts.

I suspect journal editors will fear being targeted in political witch hunts if they don’t send manuscripts concluding “lab leak” out for review and if they don’t publish them even if reviews are mediocre or worse.

For those not paying attention, scientific societies and universities are banning words and wiping web pages and records of previous programs. Zero reason imo to expect this to stop at language tangentially related to affirmative action and discrimination.

Are people here already familiar with the earlier draft of DEFUSE that provides more details on the plans to make a virus with the same features as COVID?
https://usrtk.org/covid-19-origins/scientists-proposed-making-viruses-with-unique-features-of-sars-cov-2-in-wuhan/

I think that combined with the information that EcoHealth facilitated gain of function research at WIV is pretty strong evidence:
https://oversight.house.gov/release/breaking-hhs-formally-debars-ecohealth-alliance-dr-peter-daszak-after-covid-select-reveals-pandemic-era-wrongdoing/

@DanielNadolny wait till you see the replies from the two usual suspects. 🙄

@DanielNadolny

1. The BsaI/BsmBI theory was disproven a month before the preprint came out in the same Twitter thread where most people learned about it (e.g. the same thread where Alex Washburne, an author of the paper, learned about it). The restriction sites are unambiguously natural. Found in viruses with shared ancestry. The opposite would be true for engineered sites.

2. Authors of the paper completely ignored this and continue to ignore it. They said they would address it once, but then decided to say that their answer is that all the other bat virus genomes (and pangolin virus genomes) must be fake.

3. SARS2 absolutely doesn’t match what’s proposed in this way. The shortest fragment from a BsmBI/BsaI digest is far shorter than any you’ll find in any coronavirus reverse genetics system — I checked them all. Retaining the sites in the construction like this has never been done before (someone might show up with a reference saying otherwise, but they didn’t read closely enough and are wrong).

4. The authors claim this is a perfect reverse genetics system. Yet, all around the world not a single lab used it. Instead, they made new systems. A global conspiracy of virologists to protect WIV? No, it’s just a shitty system. Prior work would suggest it will have problems because one fragment is generally unstable in plasmids in E. coli in previous work (Baric mentions this in his congressional testimony and it went over everyone’s head).

There’s a lot more that’s wrong in that article, but probably good just to focus on the frauds that Kopp and Ebright have promoted as proving a smoking gun.

Two other scientists were duped by the theory. Justin Kinney figured it out and decided to just never talk about it again and hope it goes away. Francois Balloux wrote a whole blog post about how he figured out it was wrong (in itself, kinda sad since he effectively copied someone else’s analysis and said it was a product of his own deep reflection).

@George what you don’t have any more tweeting bigots you can quote to add to the conversation?

@zcoli Thanks for replying, I'll look into that - If I'm understanding correctly, your response is against analyses based on covid and trying to determine natural vs human-mafe origins, rather than how closely covid matches what was in the earlier draft DEFUSE proposal, but please correct me if I've misunderstood.

How about the location of the furin cleavage site at the S1/S2 boundary?

@DanielNadolny My response is that SARS-CoV-2 doesn't closely match what was in the DEFUSE proposal. You've been duped by the fraud of Bruttel et al if you think so. Sorry.

Additionally, DEFUSE doesn't propose to insert random sequences at the S1/S2 site in random viruses. People are misreading this part of DEFUSE:

This part: "where clear mismatches occur, we will introduce appropriate human-specific cleavage sites." This means they'll make point mutations to what doesn't match. It doesn't mean they'll insert 4 new amino acids. For example, if they stumbled on RaTG13, they might've made HTQRNSR^STS to give the RxxR cleavage motif. If they stumbled onto SARS-CoV, they might've made HTVRLLRSTS. See this figure (borrowed from a fraudulent anti-vax paper I'm not going to link):

Here's an example of doing this at S2' for SARS-CoV - https://pubmed.ncbi.nlm.nih.gov/21435673/

A proteolytic cleavage site at S1/S2 isn't surprising. Inserts with similar lengths and compositions occur throughout the pandemic and occasionally increase viral fitness (maybe you've heard of CGG-CGG being suspicious? there've been CGG-CGG-CGG inserts elsewhere).

It's such a small insert that it will probably always be a bit of a mystery where it came from. It could've happened step by step with inserts of a few nucleotides at a time and various mutations. Mutate a single amino acid in RaTG13 and other viruses similar to SARS-CoV-2 and you get a predicted functional cleavage site that's predicted to be as good as the one in SARS-CoV-2 (probably it's not as good; predictions aren't perfect; no one's done the experiment afaik).

If you've ever heard the term "God of the gaps", you're looking at the world's tiniest God of the gaps. If you haven't heard the term, look it up and see how it's applicable to this question.

@zcoli

Look, I don't think you're really debating in good faith here, but this is a prediction market site, not a mathematical proof, and the point here is to reason from a position of uncertainty, not to prove something beyond all reasonable doubt (which we'll probably never get in this particular case anyway).

The point isn't that COVID is exactly what you'd get if you followed Defuse to the letter. The point is that if you tried to predict what researchers that accidentally released COVID would have put in a grant proposal, Defuse is about as close as what you'd conceivably imagine!

-Research labs typically only do about 1/2 of what they promise in grants, and only about 1/2 of what they do is explicitly laid out in the grant.

-Research labs typically begin new areas of research before they apply for and receive grant funding.

-Grant proposals are often rushed, edited incongruously by several people, and are weird mishmashes of different ideas. It's very unsurprising that Defuse would be quite similar to but not identical to the procedure or motivation behind gain of function research that could lead to COVID.

-Moreover, the phase space of lab leak includes cases where the virus is unmodified, in which case Defuse is from a Bayesian perspective even more likely in these worlds to be evidence for the possession of a COVID-like virus.

@DanielNadolny The restrictions sites thing didn't really get discussed in the Rootclaim debate, since Saar didn't think that paper was actually worth introducing as evidence. I did briefly respond in the written section of the debate, starting on slide 32 here:
https://drive.google.com/file/d/1N2IKOelaTz9c1unWGQ2VjXcFSwAPm5Xx/view

Yes, I included some ad hominem attacks. No, I don't regret it. Washburne and Bruttel do not conduct themselves like trustworthy people. Their theory doesn't make sense either, but understanding the science is kind of hard and understanding who's a bullshitter is comparatively easy -- if Bruttel thinks that monkeypox was a lab leak and Omicron was a lab leak, you might not want to also trust his paper on restriction sites.

When scientists did point out an obvious problem with his paper (the same sites were found in other natural viruses), he retreated to, "maybe those other viruses are all fake".

What's happened since the debate is someone found another document on the DEFUSE grant which describes making a virus in 6 pieces. And the Washburne/Bruttel theory also said 6 pieces.

Does that confirm that the theory was right all along?

I think the actual answer is that every paper Baric published before said 6 pieces, so Washburne+Bruttel made a theory with 6. And then when another document from Baric's lab saying 6 came up, they declared it an amazing coincidence.

But I'm not sure. I haven't read every paper on restriction sites, so I don't know how consistent the 6 segments thing is. Maybe @zcoli knows the answer.

I think the WIV didn't actually even use the same methods that Baric did, for making their viruses, but I also haven't dug into that topic enough to say for sure.

@benshindel I'm arguing in good faith. There's nothing in DEFUSE that leads to SARS-CoV-2. The Bruttel et al paper on BsaI/BsmBI is a fraudulent paper.

If DEFUSE happened even though it wasn't funded, that's evidence against an engineered lab leak! Because there's no path through DEFUSE to an engineered SARS-CoV-2.

If you want to argue for an unmodified lab leak, please do. It's far more likely. Just quit it with the intelligent design appeals to how impossibly unique the furin cleavage site is and so on. DEFUSE doesn't lend support for an unmodified lab leak, either. It describes work everyone already knows WIV was doing from their published work when it comes to sampling. It describes work that's less likely to have a lab leak that's successfully covered up than what WIV did previously, because constructing a small number of viruses from reverse genetics is both less risky than trying to isolate viruses from culturing a bunch of samples, it's also impossible to have a lab leak for a sequenced virus without leaving a huge paper trail that's hard to cover up.

@zcoli I appreciate you watching out regarding sources of information, though I'm a little confused - it looks like you're quoting from the submitted DEFUSE proposal, rather than the draft that's linked to in what I had shared, that seems to have some additional details and correspondence related to its development.

To clarify, I am not looking at sources that were examining the structure of covid and based on that, suggesting how plausibly it was to be constructed in part by human influence. From what I understand, the draft version of DEFUSE and surrounding correspondence is more specific there - is that not the case? If so, I really want to understand the discrepancy in what was claimed in the US Right to Know document and the draft proposal of DEFUSE I linked to, versus what is known about the structure of covid.

@PeterMillerc030 Hey Peter, thanks for replying. I'm a little confused, perhaps you can help clarify. From what I understand of this link, the draft proposal of DEFUSE, and correspondence surrounding it, Covid is in six segments, which is what the draft had proposed: https://usrtk.org/covid-19-origins/scientists-proposed-making-viruses-with-unique-features-of-sars-cov-2-in-wuhan/

You mentioned Baric - he is one of the collaborators with correspondence in that draft proposal.

That draft also includes figure C on page 524:

The US Right to Know article (which includes a link to the draft proposal and correspondence I'm talking about), claims that this is showing a furin cleavage site in the same place it's found in Covid.

This draft proposal and documentation seem critical for understanding the potential links between the proposed work involving WIV and Covid - is this document something you haven't yet had the chance to explore?

@zcoli you seem to be overindexing on specific technical details in the Defuse proposal.

DEFUSE proposed that a lab in Wuhan should work on gain-of-function work on bat coronaviruses to make them more transmissible between humans by inserting a furin cleavage site, in 2018-2019

We know that COVID looks like a bat coronavirus that is markedly more transmissible between humans due to a furin cleavage site, and that it emerged in 2019 in Wuhan.

The proposal from Defuse is the closest single grant proposal (that I'm aware of) to what we'd expect a grant proposal to look like at a lab where covid emerged. If you have a counterexample of another grant proposal that would be closer, feel free to share it.

If that isn't massive support from a Bayesian perspective, I don't know what is.

@DanielNadolny It's a confusing subject. I might recommend watching:
https://www.youtube.com/watch?v=FLnXVflOjMo

Download the slides, pursue various links -- read more about what furin cleavage sites are, what they do, read about previous experiments that modified one. Maybe also watch Yuri's presentation and our debate. Read and fact check as much as possible for both sides. That's a rabbit hole that will last you weeks/months, if you really want to get into this -- I was undecided on covid origins until I had spent ~3 months researching it. I gradually realized that almost all of the case for zoonosis holds up, and almost none of the lab leak stuff does.

For a shorter summary:

I think that the Washburne+Bruttel 6 segments paper does not hold up in any way.

The presence of a furin cleavage site in SARS-CoV-2 is actually interesting. That's been found in some bat viruses and other human viruses, but it has not been found specifically in other bat sarbecoviruses (out of about 1,000, I think? or that order of magnitude).

There are basically 3 options here:

1. some bat sarbecoviruses have a FCS, but it's very rare and hard to find a virus like that.

2. bat sarbecoviruses never have a FCS, but that can evolve after the virus has spilled over to an intermediate species (these are intestinal viruses in bats, but they become respiratory viruses in other species, so evolution starts to favor having the FCS)

3. scientists engineered that site into the virus.

I favor explanation 2. Most lab leakers prefer 3, a handful of scientists prefer explanation 1.

I think the FCS is basically 1 of 2 features of the lab leak theory that holds up to scrutiny (the other 1 being that there is a virology lab in Wuhan). Just about everything else about the lab leak theory falls apart, under scrutiny: the timing of the outbreak, the locations of the cases, the "3 sick researchers at WIV", the double CGG, etc.

The location of the FCS at S1/S2 is basically meaningless as evidence. There are only 2 places where you would put a furin cleavage site -- either S1/S2 or S2'. The final DEFUSE grant seems to say they were interested in S2', and some virologists who favor zoonosis said that is evidence that DEFUSE didn't make covid. Some people who favor lab leak read the new draft as interested in either spot. I think both arguments are basically worthless, because you'd be talking about odds of 1 in 2, either way.

There are several reasons why the specific FCS found in SARS-CoV-2 does not look engineered:
1. it's out of frame, when an engineer would prefer to put it in frame

2. it's inefficient (engineers prefer efficient cleavage sites like RKRR or RRSRR, while Covid uses PRRAR).
   No virology experiment in history has used an inefficient choice like that.
   The proline (P) in particular is a bizarre engineering choice that scientists would be unlikely to make. After covid spilled over, the P evolved into better choices like H and R.

3. it's an insertion, not a mutation. No virology experiment in history has added a furin cleavage site via insertion.

And there are other important ways in which SARS-CoV-2 does not match what was written in the DEFUSE grant. The DEFUSE grant says they're going to put novel spikes into known backbones (like WIV1 and SHC014). SARS-CoV-2 does not have one of those backbones. If it did, everyone would agree it's a lab leak.

The lab leak scenario right now is basically:

1. WIV secretly discovers a novel virus with the right RBD, but for some reason does not publish it along with the rest of their sarbecoviruses in 2018 and 2019.

2. The DEFUSE grant is cancelled in the US but happens in Wuhan instead.

3. WIV makes a novel reverse genetics system for that virus, and uses that instead of using the existing backbones that DEFUSE said they would use.

4. WIV adds an inefficient FCS sequence that's never been used before, via insertion which has also never been used before.

5. WIV makes additional modifications like N-glycan changes that simulate natural evolution in an intermediate host

6. WIV finds some way to grow this virus without leaving a genetic trace (most cell cultures would)

7. The virus infects someone at the lab, who immediately goes across town to a wholesale market and coughs on someone in a raccoon dog shop

8. A week later, a second person from the lab also goes across town to the same market and gets someone else at that market sick.

9. The virus causes no other case cluster in Wuhan or at the lab.

10. There's a perfect cover up, and no one ever finds details on that secret starting virus, the sampling trips to find it, the lab worker who got sick, etc.

None of that is impossible, but there are many highly unlikely steps involved, and the competing theory is just: it was sick animals at a market, the same way SARS1 started, and one shop was selling the exact same animals that started SARS1. My money is on the sick animals theory.

@benshindel As of 2020, one of the most popular lab leak theories was that miners caught covid in a cave in 2013, Covid adapted naturally to humans in their lungs, WIV took samples and froze them, then WIV leaked those samples while experimenting with them in 2019.

People said it couldn't be a coincidence that WIV had been to that cave, that those miners had gotten some respiratory illness, and that WIV had discovered a virus 96% similar to Covid in that cave.

Today, most people have moved on from the Mojiang miners theory, and many now say that the DEFUSE grant can't be a coincidence.

It's pretty easy to read too much into the details of various kinds of research, and to note similarities that seem interesting.

If you want examples of people reading too much into different research grants, go read Jim Haslam's work on Twitter or Substack, or buy his book.

He argues that Covid was not made by DEFUSE but by other grants done by guys like Baric and Munster in the US, they created a new virus, then they exported the work to Wuhan via the single foreign scientist who was working over there (Danielle Andersen, at the Wuhan BSL-4). She leaked it by accident and Chinese scientists were never involved in the creation.

He makes it sound just as convincing as the theories where Chinese scientists and the DEFUSE grant are to blame. Steve Hsu recently had Haslam on his podcast, and Steve seemed to find it all plausible.

@benshindel I’m focusing on the part of DEFUSE that doesn’t suggest “inserting” a furin cleavage site in the way you say it does. I showed that exact part of the proposal, explained what it means, and gave a reference for an example of what it means in the literature.

The fantasy interpretation you believe in has no pre pandemic precedent in the literature because no one would do an experiment that way and DEFUSE does not propose to do it.

“The closest proposal” doesn’t help you if it’s still infinitely far away.

@DanielNadolny USRTK lies about the relationship between DEFUSE and SARS2. The narrative they go with, led by Emily Kopp, is the 100% false narrative of the BsaI/BsmBI theory. They lie about where one of these two enzymes shows up in a grant proposal.

The discrepancy is that you’re trying to rationalize people who lie to profit off of conspiracy theories making villains out of random scientists and the truth. Go read their attack on Peter Hotez and do your own research to see how BS it is. Then stop wondering why it’s hard to square what you read at USRTK with reality.

@PeterMillerc030

People said it couldn't be a coincidence that WIV had been to that cave, that those miners had gotten some respiratory illness, and that WIV had discovered a virus 96% similar to Covid in that cave.

Today, most people have moved on from the Mojiang miners theory, and many now say that the DEFUSE grant can't be a coincidence.

This is an important point. The first burst of engineering-based lab leak theories were:

• Li-Meng Yan report: cherry picking restriction sites from ZC45/ZXC21 viruses sampled in Zhejiang province and I don't recall the specifics of what else is spliced other than that it was clearly not worth committing to memory

• RaTG13-or-similar adapted for humans via some mix of the following:

  • Selection human cell culture and/or genetically modified mice

  • Rational genetic engineering to grab a polybasic cleavage site at S1/S2 from somewhere or other. The quantity of proposed sources out there (one person's HIV sequence, the human ENaC gene, one cat's feline coronavirus sequence, a nuclear localization sequence) proves that the specific theories are all meaningless overfitting

  • Engineering the receptor binding domain (RBD) either rationally or by splicing in a sequence from a pangolin coronavirus

The receptor binding domain part of the theory was absolutely essential. This was a much larger gap in the evolutionary record than the S1/S2 cleavage site. I don't think I saw an engineering theory of SARS-CoV-2 origins that assumed the RBD was natural. That, was, until the BANAL-20-52 genome was published, instantaneously disproving all of the lab leak theories that assumed an artificial RBD.

Losing the strongest molecular evidence that SARS-CoV-2 defied an explanation in nature should've dramatically lowered everyone's odds for a lab leak origin of an engineered virus. Instead, people just quickly came up with new theories and said they were even more certain now... beats me.

@DanielNadolny Baric in his testimony describes how those six segments include one that is very small and not what a designer would choose. In comments to the bioengineering paper, there is another comment which makes the same point. http://disq.us/p/2rjc5fr The entire comments section for the paper are worth reading, particularly the Crits-Christoph thread, the comment by disqus and the one by Jared Roach that maximum fragment size is not a robust statistic. Note how when confronted with similar sites being present in nature, the author switches to calling the other published genomes also part of a cover-up.

@benshindel This is what a bad faith argument looks like. The arguments for why Covid is most likely of natural origin are technical and one cannot gloss over the fine details in favor of circumstantial pattern matching.

@BW Hi BW, I just read through that testimony, thank you for bringing my attention to it. There were quite a few things I found very interesting in it.

Checking in with @PeterMillerc030 and @zcoli, are you aware that Dr. Baric says in his testimony that Covid emerged in the middle of October or later October?

It's on page 101: https://oversight.house.gov/wp-content/uploads/2024/04/Baric-TI-Transcript.pdf

There's a lot of other interesting information shared in there - if you haven't yet read it, let me know and I can share some of what's discussed!

@PeterMillerc030
Thanks Peter! I think it's clear I do have more research to do, and I'll keep at it.

For reference, here's the relevant part of that draft proposal and communication I mentioned. It's on page 134 - I highlighted the relevant section. It says that Baric would reverse engineer spike proteins in his lab, which would then be inserted into SARS-CoV backbones, and inoculated into humanized mice - there's a comment too that mentioned the work would not necessarily be done by doctor Baric, but a lot of the assays could be done in Wuhan. I'm interested in what you make of it - if this isn't relevant, why not?

Also, I mentioned in an above comment that Dr. Baric in his testimony says that based on the molecular clock, Covid emerged in October, and that the first known case was five or six transmission cycles down the line.

@DanielNadolny I have read Baric's testimony. I don't know why he said October. I don't think anyone knows, because he's never said or written anything about it.

Every reasonable model of the pandemic that I'm familiar with says it probably started in late November. See, for instance, Pekar et al 2021, Pekar et al 2022, Jijon et al 2024. I've also tried simulating it myself, and wrote up some results on Twitter:
https://x.com/tgof137/status/1772417277670871113

It is theoretically possible to have a few cases going back as far as October, but it's a low probability thing:

https://x.com/tgof137/status/1772417301561708753

The "molecular clock" quote doesn't make any sense to me. That just sounds like he's talking about genetic TMRCA, which for covid is early December, a little bit later than the actual first case.

I'd love to talk to Baric and understand why he disagrees, but I don't know him.

It is amusing how the lab leak movement has picked up on that quote to say that the market origin theory is disproven. Like, elsewhere in his testimony, Baric said that the Washburne/Bruttel paper was trash, but they don't take his word for that. And many of the lab leakers think he wrote the DEFUSE grant and that grant created SARS2. Some of them (like Robert Redfield) actually think that Baric made the virus himself.

So if virologists were all trying to lie and cover up the lab leak, wouldn't Baric be the first person to jump aboard the market origin theory and support the cover up?

I tend to think it makes more sense to listen to what scientists share and publish, i.e. if someone writes papers about when the pandemic started, I read those. Baric doesn't study that -- he writes papers about how he makes viruses, so I read those.

@DanielNadolny First off, Covid does not use a SARS-CoV backbone. When you say SARS-CoV backbone, you'd be referring to WIV1, WIV16, and SHC014. Those are the 3 that the WIV had created before and knew how to work with. The DEFUSE grant says that they're going to put new spikes into those backbones and see what happens. We know for sure that's not how Covid was made. If it was made that way, you'd be able to easily see that -- it would be identical to one of those known viruses everywhere outside the spike.

Covid instead looks like a totally new virus, no part of it matches any published virus.

So, even if you did create such viruses in Wuhan, and you infected mice with them, and someone caught it from a mouse, or whatever, it would not make Covid (it would be a different, recognizable virus).

Second, I think the comment in the margins is not actually referring to the whole paragraph, but to the subset of it that I've marked in red:

And I think that's the case, because we have the final grant, and that already described doing the viral binding assays at WIV:

So the draft and the final version actually say the same thing! But lab leak supporters quoted the draft to say we had discovered something shocking and new. This is one of the many ways that the lab leak people manufacture controversy.

Pseudovirus binding assays could not create a pandemic, even with horrible lab safety -- pseudoviruses can't replicate.

I think live virus binding assays could cause an accident, but again, if such a virus got out, it would not be covid. They'd be creating it with a known backbone, because at this step they'd just be screening different spikes to find out which ones can infect human cells. They also wouldn't have gotten around to messing with the furin cleavage site, yet.

@DanielNadolny As stated by @PeterMillerc030 and @Z , the natural origins argument is a probabilistic argument, just not a 'common sense' probabilistic argument that stops with a high-level inference based on the Wuhan location and DEFUSE. But one that gets into the details of what exactly claims of an engineered virus involve or how one could get from DEFUSE to SARS2. At first, it was a pangolin virus chimera to explain the RBD. When that failed, it was discarded in favor of reverse genetics for the FCS. But the natural origins argument is that the reverse genetics would have to be one that an engineer would not choose and that the claimed fingerprints are false. The FCS is one that an engineer would not choose i.e. non-canonical, out of frame, suboptimal etc. There is no way to gloss over all these details when arguing for an engineered virus. When people claim FBI, CIA scientists are on the lab leak side, for what it's worth (and it's not much) they are also ignoring the earlier intelligence report that almost all intelligence agencies do not believe that the virus was bio-engineered. So the case for lab leak generally involves data dredging and discarding or using pieces of evidence where useful to advance the case. Baric is useful when he says that the market was not the origin (when epidemiology is not his field of expertise) but not when he says that the virus was not engineered (on which he is indeed an authority). If you believe that the pathway to SARS2 involves bio-engineering, the comments to the Bruttel paper are very pertinent in arguing for natural origin via recombination as being the more likely pathway.

@PeterMillerc030

You say stuff like:

"The lab leak scenario right now is basically:

1. WIV secretly discovers a novel virus with the right RBD, but for some reason does not publish it along with the rest of their sarbecoviruses in 2018 and 2019.

2. The DEFUSE grant is cancelled in the US but happens in Wuhan instead.

3. WIV makes a novel reverse genetics system for that virus, and uses that instead of using the existing backbones that DEFUSE said they would use.

4. WIV adds an inefficient FCS sequence that's never been used before, via insertion which has also never been used before.

5. WIV makes additional modifications like N-glycan changes that simulate natural evolution in an intermediate host

6. WIV finds some way to grow this virus without leaving a genetic trace (most cell cultures would)

7. The virus infects someone at the lab, who immediately goes across town to a wholesale market and coughs on someone in a raccoon dog shop

8. A week later, a second person from the lab also goes across town to the same market and gets someone else at that market sick.

9. The virus causes no other case cluster in Wuhan or at the lab.

10. There's a perfect cover up, and no one ever finds details on that secret starting virus, the sampling trips to find it, the lab worker who got sick, etc."

This kind of rhetoric makes me want to never take your claims seriously because I suspect you are arguing from bad faith. If you divide any event into infinitely recursive sub-events and assign probabilities to each of them, you'll make the likelihood of any event rapidly approach zero.

But also, basically NONE of these steps are necessary for the lab leak to be plausible or even likely! The fact that you believe or want others to believe this is very suspicious, and it's the same manner in which you argued during the rootclaims debate, when you assigned absurdly overconfident Bayes factors to a bunch of hyper-specified criteria in an intellectually dishonest way. In that debate, you were at least doing it for a financial reward. But here you're just mucking up the discourse in a way that's epistemically damaging to this site's users.

I can go through these one by one.

1. Ya, this makes total sense and is more likely than the opposite. Labs can take years to publish data and probably only rigorously document like 10% of what they do.

2. This doesn't make sense, what are you even arguing? That if lab leak were true we'd see simultaneous COVID releases in US and Wuhan? Think about if this makes ANY sense.

3. Research basically never follows ever step described by grant proposals. Moreover, DEFUSE is just an update towards the lab doing activities that are very similar to those that might create a virus like COVID (as I explained earlier in the thread). You have to basically have zero understanding of how STEM labs operate if you think that grant proposals are useful in establishing a precise sequence of research activity.

4. Not necessary. Lab researchers could have been fucking around testing a process. There could have been errors induced in the process. The viruses could have mutated within media, animals, humans, before being detected. There is a massive phase space in which this does not need to happen.

5. See above.

6. What are you talking about "Genetic Trace"? Have you ever even SEEN a wet lab? What level of diligence and detective work would it have required to upturn something like this? Especially without explicit cooperation from whichever unfortunate scientist might have accidentally caused the outbreak?!

7. There's no reason to believe that the first case, vs the first major spreading event was at the wet market. It could have been a chain of several ppl leading to a spreading event at the wet market. But it's also the closest market to the CDC facility, and many ppl could plausibly have had meetings/work/appointments at both sites.

8. This chain of events is not established for natural release either. You're just making this shit up!

9. There were tons of other cases across Wuhan during this time period. See above.

10. Not required in the slightest!

@benshindel I've never seen someone get so triggered by an enumerated list.

Do you also react in horror when Alina Chan writes "Why the Pandemic Probably Started in a Lab, in 5 Key Points"?

When ChatGPT responds to a question with several numbered items, do you accuse OpenAI of bad faith?

Unfortunately, you don't seem to have any understanding of most of the points I made. For instance, in step #6, "a genetic trace" has nothing to do with lab contamination. I am referring to a mutation that would happen in cell culture. For instance, if you culture SARS-CoV-2 in Vero cells, it loses the furin cleavage site. Therefore, it wasn't made in Vero cells. If you culture it in Calu-3 cells cells, mutations occur in the envelope protein. Those were not seen in early strains of Covid, so it was not cultured in Calu-3 cells. I spent days, along with my research assistant (Shin Jie Yong), reading every paper we could find on culturing of SARS-CoV-2 to see if we could find any pathway that would work here, made numerous slides, and presented them in the second debate session.

My goal with this whole project has been to help educate people about the science on an issue that's been very poorly represented in most media (lab leak is exciting and sells well, zoonosis is boring yet almost certainly true). Many people have found the effort valuable.

I'm sorry that it hurts your feels.

@PeterMillerc030

I love numbered lists. It's not the numbered list, it's the way you frame it as requiring each of these things in succession!

I misunderstood your use of the term "genetic trace". I think you're dramatically under-estimating how poorly calibrated most scientists (including geneticists) are about out-of-model and out-of-domain uncertainty. For example, when you say that "you couldn't find a single pathway that would work here" you seem to be failing to consider several compounding levels of uncertainty:

Assuming lab leak,

-What are the odds that the virus leaked from the lab unmodified?

-What are the odds that the virus mutated enough to confuse this analysis between the initial generation in the lab and first detection?

-What are the odds that any of the pathways you read about was misrepresented or explained inadequately in the paper, such to confuse the analysis?

-What are the odds that the virus was generated in a pathway you didn't read about?

etc

What's your true forecast on this market? How confident are you truly? 75%? 90%? 99%? 99.9%?

Because based on your Bayesian analysis at the Rootclaim debate, one might think you have like 10 sigma confidence

@PeterMillerc030 But also, there's a massively disproportionate burden of research here.

If I'm arguing with a UFO truther, and I say, "well a lot of ppl think that photo is just a high-altitude balloon," the UFO truther will respond:

"Ah well, my assistant and me found every single patent for a high-altitude balloon and not a single one matches the geometry that's visible in this photo."

I mean, this is hard to counter. I personally have not reviewed every high-altitude balloon design, nor do I wish to spend dozens of hours doing so. But generally, I can imagine that

-The balloon might have become deformed

-There could be balloon designs that weren't patented

-The imaging conditions might have introduced an optical artifact in the image

It's frustrating that you and Z seem to be engaging in this sort of asymmetrical debate where rather than evaluating the strongest cases from both natural release and lab leakers, you instead try to strawman lab leak arguments based on the most brain-wormed internet commentators

@benshindel I don't think I engage with only the weakest lab leak theories. I bet $100,000 against the lab leak theory and debated against the co-founder of DRASTIC and some multi-millionaire who hired a whole research team. If they didn't have a good theory, I don't think there is some strong version of the lab leak theory.

What's my confidence for covid being natural, accounting for out of model error? I'm happy to say it's over 99%. I don't really see a ton of value in arguing over decimal points beyond that. Lab leak is certainly not 1 in 10^26 likely, or whatever that heavily criticized slide said, nor is there any meaningful way to think about numbers that small. As I said at the debate, I think that bayesian calculations are often just set up to give some illusion of objectivity.

What's my confidence in this market ever resolving? Relatively low. We have equivalent markets for Lyme disease and HIV, sitting at 11% and 8%. You could make a market for "is your mom a lab leak?" and that might crack double digits. I don't know how we get to 98% agreement, or whatever the requirement here is.

We have some markets on whether the resolution will be controversial. I think it would be uncontroversial if Covid is proved to be a lab leak and controversial if Covid is proven to be natural. I think scientists are rational people that would update upon seeing convincing evidence of a lab leak. I think many of the famous lab leakers are not rational or honest people and would never update on any amount of evidence for zoonosis. I can't think of any piece of evidence that would convince all of them. I can think of many things that would change my mind.

I suppose I should point out that many of the people who are on the sidelines, or weakly leaning one way, are rational and reasonable. But I think that there will still be a core group of lab leak promoters who make dishonest arguments that sound persuasive if you haven't thought hard enough about it. They will always have more followers, be more prolific, get more media attention, and so on. Scientists will keep trying to rebut claims via careful published research, spending a year to analyze every small issue just to write a paper that few people read. Lab leakers will continue to make explosive claims via social media that spread much more rapidly.

What was my confidence level for the debate? Emotionally, it was scary as hell to bet that much money, with no prior debate experience, no knowledge of Saar's skill, arguments, or tactics, and some vague concerns on if it could be rigged. But, as I said elsewhere, my logical guess was that intelligent judges would each vote for natural origin at a 80/20 rate. I said that because the evidence for a natural origin is much better, but some people still fixate on the lab leak theory. If the 2 decisions were uncorrelated, I figured 4% of severe financial loss, with a higher risk of it going split decision. There are of course also various correlated ways that I could lose, like if I choked at the live debate.

After the debate, I jotted down a rough guess at the odds of the 6 outcomes (win/lose/draw) for each judge. I have it written down somewhere, I could check. IIRC it was actually pretty close to what the markets were pricing, post debate videos and pre-decision. Maybe that makes me well calibrated. But Saar was also certain he had won, and sent me his victory announcement to proofread.

I would bet against lab leak again today, given some kind of non-debate means of resolution (i.e. a wait a few years to see how the evidence unfolds kind of thing). The odds of winning that would be above 50%. If the actual odds of lab leak are < 1%, then the odds of losing such a bet are definitely higher than 1%, because the judging mechanism is not guaranteed to get the correct result. I think the biggest risk involved at the moment would be someone in the Trump administration/intelligence community just making up evidence in support of lab leak. Kash Patel and John Ratcliffe are Trump loyalists and not honest guys.

I'm unsure if I would do the debate thing again. I think the odds are better than 50% that I would win again, but they're not > 99%. It was stressful, it took more time than I expected, I can make more money in the same amount of time writing code. I already won over a lot of the people that I was hoping to persuade. And the evidence base has not changed much since the first debate, so it would be a lot of repetition for the same result. I suppose that the more time elapses, the more that latter point could change.

@DanielNadolny Re: Baric and October

This is a common-sense extrapolation if you include as a data point and SCMP article that claimed to reveal some confirmed cases in November. See also this 2021 article coming to a similar conclusion with data and modeling and not gut feelings: https://www.science.org/doi/10.1126/science.abf8003

It's also the wrong conclusion, which is why the authors of that paper wrote another paper that did not include the SCMP "data" and came to a conclusion of mid-November for the first spillover that sustained transmission. But a spillover sometime in October is still plausible and consistent with the first ascertained cases having onset in the first half of December.

The reality is that the SCMP article was talking about cases that had some clinical COVID-19 features (they'd been "confirmed" to match some clinical parameters after being initially flagged by clinics) but they had not and were apparently never "confirmed" by molecular diagnosis. The evidence for this is (1) the cases are randomly distributed rather than increasing over time [see the WHO report, where a time series pretty closely matching what's described is shown], (2) at the same time the retrospective case search was reported to SCMP, it was also reported in Health Times (a publication from Peoples' Daily) with similar features.

Imagine there's a cover-up going on in late February 2020. Don't you think that Peoples' Daily is gonna be on the ball when it comes to not publishing covered up information?

So, the explanation for this is simply that Baric hasn't read literature published since 2021 closely because he doesn't think it's worth the time. There are other parts of his interview that are also consistent with this.

@PeterMillerc030 @zcoli

The molecular clock has to do with the mutation errors that come with replication. It looks like most papers using this kind of analysis suggest an earlier emergence for covid. This kind of approach doesn't have the same problems epidiomologicial approaches have, due to asymptotic carriers.

Even a few one-to-one transmissions is consistent with how covid spread throughout the pandemic, with many people infecting 0 or 1 other person. From what I understand, Peter, you already acknowledge that the first known case is not actually the first human case.

Peter, you gave a date of December based on molecular evidence - can you share some sources there?

Z, just to make sure we're talking about the same Ralph Baric, this is the person who is one of the world wide experts, who has in-depth molecular knowledge having done work that plausibly raised concerns he could have created covid. You're suggesting that when he said based on the molecular clock there was no chance the first known cases were right at the start during his testimony, he decided to potentially perjure himself in his own area of expertise, based on his opinion not agreeing with yours?

@PeterMillerc030 Hey Peter, if you look at the snipped clip I shared right before the part I highlighted, it says "Isolation will be attempted on a subset of samples with novel SARSr-CoVs" . You pointing out that Covid is a new virus is consistent with that proposal, not refuting it.

Based on that side comment and other correspondence, we know the people writing the proposal were worried about the red flags doing the work in China would raise, and intentionally downplayed it. You might recall this was one of the topics that came up in Baric's testimony - his warnings and concerns with the level of biosafety practiced in the Wuhan lab.

@DanielNadolny You don't need DEFUSE to argue that Zheng-Li Shi's research program involved work with isolated and recombinant SARS-like coronaviruses. What you learn from DEFUSE is that Zheng-Li Shi and others were planning a collaboration where coronaviruses would be identified in sequences from samples collected in China and sequenced in Wuhan, with subsequent steps being shared by partners all around the world, subject to approval by ethical boards around the world, and done at BSL-3 rather than BSL-2.

If you step back and try to forget the alluring conspiracy theory about DEFUSE being covered up until bravely rescued from a secret server and leaked by a whistleblower (the IG investigated that, and it wasn't true), you see that, if you think DEFUSE happened anyway even if the grant was rejected, the addition of DEFUSE makes a lab leak less likely than if DEFUSE hadn't existed. Zheng-Li Shi's lab was already sampling from bats and growing and modifying viruses; that's not new. What's new is adding a bunch of partners and shifting to a sequencing-first approach that means a bigger trail of evidence and more people to keep quiet to succeed in any cover up of a lab leak. Bonus: the addition of Baric convinced the team to propose work at BSL-3 rather than BSL-2; work at WIV at BSL-2 is often cited as evidence of lab leak (this replaced previous evidence of lab leak: the existence of a facility for BSL-4 work in Wuhan).

And, yes, I know Emily Kopp at USRTK invented a conspiracy theory in which all of the collaborators secretly plotted for WIV to do work at BSL-2 by reading something absurd into someone's track changes history.

@zcoli Thanks for continuing the discussion.

What we learn from that draft proposal is that the specifics of the proposed work closely match with Covid. That doesn't require any argument from Emily Kopp.

Let's go over the argument here a bit more, as I think it's getting lost in the fragmenting of the discussion.

There was a proposal to add a furin cleavage site to the location it is in Covid, that it was to be done on a subset of novel Sars-Cov backbones, that they'd use humanized mice, that some of that work could be done in WIV, and that we know WIV was doing gain of function research without appropriate levels of safety precautions.

On the other side, there is an absence of good evidence for an animal vector that brought covid to wet market (though there are possible candidates), and with five or six transmissions prior to discovery (according to expert testimony), the wet market location doesn't really show anything aside from being a good site to spread covid around.

I can appreciate that people have strong Priors here, but the string of unlikely events that would have to occur that Peter had laid out are largely things that are consistent for what was proposed or that we know about.

On the other hand, the relevance of the wet market hinges on the first known case to be almost at the start of the pandemic. People infected with covid often spread it to 0 or 1 other person, so the plausibility that it could have been transmitted one to one for some generations before spreading widely is already high, and on top of that we have expert testimony that puts the first known case five or six transmissions down the line, suggesting the only thing the wet market really showed is that it's a good place to spread covid.

The zoonotic origins argument hinges on evidence that is far weaker than you or Peter seem to believe, and meanwhile the unlikely things that would need to be true for a lab leak are so close to what is proposed, it acts as evidence for the lab leak, not against it.

@DanielNadolny Baric isn't an unimpeachable authority on everything and this isn't his specific area of expertise. I don't blindly trust him, either, when he says, "it's more likely to be a natural event than it is to come out of the laboratory. The data -- that's what the data screams."

Baric said: "If you look at the molecular clock of the virus, it emerged in the middle of October, late October, not the middle or end of November."

You cannot say this from "molecular clock" analysis. You must combine some type of molecular clock analysis with some model of disease transmission and the likelihood that an infected person leads to a sequenced virus. You probably want to account for that likelihood being time dependent and biased ascertainment from clustered case detection (and, sometimes, multiple sequences from the same patient).

There are three ways to get to dates of first human infections prior to November:

1. Late October is perfectly plausible (it's in the 95% hypothesis interval in Pekar 2022) and it's odd Baric sees a disagreement of a couple weeks as such a big deal in the first place. It's also got nothing to do with lab leak vs zoonosis. Lineage A and lineage B are still both in Huanan market and nowhere else first.

2. You can push back estimates if models assume reports of suspected cases in November (SCMP, Health Times) were reporting COVID-19; obviously the range of plausible dates of first infections doesn't include days after the first infection. Almost everyone misses that this is the major difference between Pekar 2021 and Pekar 2022 when it comes to the estimate of the first infection date; not multiple spillovers. Without including that report, it's 17-Nov in the 2021 paper vs 18-Nov in the 2022 paper.

3. You can push back estimates if you assume (1) a single spillover and (2) the first human infection had a mutation towards viruses sampled in bats. There isn't strong evidence to support this and there are multiple candidates for what this mutation could be. If one of the early mutations towards bat viruses is a coincidence, the others could be, too.

@DanielNadolny

The zoonotic origins argument hinges on evidence that is far weaker than you or Peter seem to believe, and meanwhile the unlikely things that would need to be true for a lab leak are so close to what is proposed, it acts as evidence for the lab leak, not against it.

Here's the first statement of the Huanan market hypothesis that I'm aware of; posted to Weibo (Google translated here) on 30-Dec-2019.

This was (and still is) just a random concerned citizen in Wuhan. It was covered in the media and that person is responsible for the media showing up and investigating the market and observing inspections of it on 31-Dec.

All that was known then was social media rumors of something SARS-like and a market-linked cluster of pneumonia.

Did the data support or reject that hypothesis subsequently?

1. All genomic diversity was found in Huanan market environmental samples

2. Retrospective searches found COVID-19 with December 2019 onset spread from the area around Huanan market. Additional data is consistent with this genuinely reflecting the distribution of infections and not ascertainment bias (healthcare outbreaks, excess deaths, data on January cases)

3. COVID-19 cases and SARS-CoV-2 samples in environmental testing were in the same corner of the market as animals that can be infected by an transmit SARS-CoV-2

4. Sampling of related viruses in bats and pangolins shows the genome of the virus is consist with expectations from viruses sampled in the wild and in trafficked wildlife

That's the evidence in a nutshell. What's the weakest link? I can't find any evidence that's gone against this hypothesis.

The first time a lab leak was suggested was not long after. The first time it showed up in the media that I know about is this article in Radio Free Asia: https://www.rfa.org/english/news/china/wuhan-outbreak-01092020133656.html

"It's a new type of mutant coronavirus," Ren said. "They haven't made public the genetic sequence, because it is highly contagious. From what I can tell, the patients caught it from other people. I have thought that all along."

She said the lack of fatalities didn't indicate that the virus was less deadly than SARS, just that antiviral medications have improved in the past 10 years or so.

Ren said she also regarded the relatively high number of infections in Hong Kong with suspicion, given that there had been no reports of cases anywhere in between the two cities, in the southern province of Guangdong, for example.

"Genetic engineering technology has gotten to such a point now, and Wuhan is home to a viral research center that is under the aegis of the China Academy of Sciences, which is the highest level of research facility in China," she said.

Repeated calls to various numbers listed for the Wuhan Institute of Virology under the Chinese Academy of Sciences rang unanswered.

Retrospectively, we know that some of this was wrong: the Chinese CDC wasn't hiding the genetic sequence because it was too dangerous; they had already submitted the sequence to GenBank in the USA in December.

I don't know of any subsequent evidence in the opposite direction. I get that you find the absence of sampling the SARS2 S1/S2 sequence in animals in nature to be very strong evidence. If you spent more time comparing the SARS2 sequence to others you could find additional gaps in the evolutionary record.

@benshindel The CDC location near the market was operational only in December 2019. Both lab leak and natural origins agree that the spillover occurred before December.

@DanielNadolny DEFUSE did not happen in a vacuum. Set aside that it was not funded. The proposal reflects a knowledge of coronaviruses over decades, what makes them less or more infectious etc. Pekar constructed a recombinant common ancestor that is 98.8% of SARS2 similar to the 98.6% recCA of SARS1. Originally the RBD was seen as a sign of genetic engineering until it was discovered that exactly the same can be found in nature. So, it boils down to the 1% that similar to SARS1 could have occurred via recombination and from time spent in an intermediate host. We are down to the FCS which is a first in sarbecoviruses and the lab leak claim is that this was inserted via reverse genetics. You brought up the six pieces and I pointed you to why in peer review, this was seen as weak. This includes nominally pro-lab leak scientists such as Alina Chan, Francois Balloux, Justin Kinney and lab leak-agnostic scientists such as Michael Eisen. I don't think you want to engage with this. Baric might think molecular clock suggests October but as others have pointed out, this is not a strong claim. The 95% range in Pekar'22 does include October anyway. In any case, as you must know since you seem to have spent some time on this issue already, the lab leak claim links the leak to a September incident when the database was taken down. October spillover, however unlikely, still doesn't get you there.

@BW There was a thread on X once where somebody listed all of the work that was done in prior decades and was reflected in DEFUSE.

https://x.com/BexZex/status/1789939478255051047

@zcoli I'm going to respond to both of your comments here to keep things more focussed.

First of all, can you provide some sources suggesting the shortcomings of molecular analysis that you're suggesting here? This is a technique that is widely used.

Secondly, you mention this:

You can push back estimates if you assume (1) a single spillover and (2) the first human infection had a mutation towards viruses sampled in bats. There isn't strong evidence to support this and there are multiple candidates for what this mutation could be. If one of the early mutations towards bat viruses is a coincidence, the others could be, too.

Have you already read "Evolutionary trajectory of diverse SARS-CoV-2 variants at the beginning of COVID-19 outbreak"?
https://academic.oup.com/ve/article/10/1/veae020/7619252

Here's their conclusion: In sum, although multiple lineages of SARS-CoV-2 were co-circulating during the early period of the COVID-19 epidemic, they still exhibited the evolutionary continuity. All of them may have evolved from one common ancestor, probably lineage A0 or an unidentified close relative, and jumped into human via a single zoonotic event. Various mutations have driven the rapid diversification of SARS-CoV-2, with some being beneficial for its better adaptation and circulation in humans, which may have determined the waxing and waning of various lineages.

In short, the best evidence right now suggests a single zoonotic event, which suggests the earlier estimates for the the origin of covid spreading people are better, given what you shared.

Regarding your second comment, I'm extremely confused. I don't see the relevance of a particular argument starting in a specific time, place, or group - I think we're here to discuss the evidence that exists for and against the possible origins of Covid using the best knowledge we have available now. Moreover, it sounds like there was early disagreement amongst experts in terms of whether covid could be a lab leak - I'm not sure why you're highlighting the first time the idea appeared in the media as some kind of evidence for or against a lab leak.

@BW Hi BW, can you explain more why you think pre-existing work that informed the DEFUSE proposal makes a lab-leak less likely?

Just to note too - I haven't actually brought up the database going down. Perhaps that's something we could discuss more broadly, but for now, it doesn't seem to be something essential for a lab leak to be likely.

What does seem relevant is that WIV apparently had many bat samples not yet worked out or published: https://usrtk.org/covid-19-origins/wuhan-institute-of-virology-has-many-unreported-bat-virus-samples-collaborating-virologist-says/

@DanielNadolny You seemed to think that inserting an FCS in the location that it was found in SARS2 makes lab leak more likely. As far as I know, DEFUSE did not say S1/S2 and said S2' but maybe there was something in the drafts? In any case, prior work understanding the role of FCS at S1/S2 does not make lab leak less likely but it does not make lab leak more likely which is what I understood you to be saying. WIV has been consistent with how many sarbecoviruses that it had samples of. DEFUSE said >180 and they published ~220.

I notice that you are still not engaging with the reverse genetics and the six pieces paper.

@DanielNadolny On bsky, you will find a thread where Kristian Andersen discusses this paper. Bloom was a proponent of A0 as ancestral, but even he backed off grudgingly on this. In short, Lv et al does not correct for the various software anomalies that Pekar '22 accounts for. There were one or two TTs that Andersen did identify which could have occurred via reversion given that the TT occurred late and there were a large number of cases. I understand him to be saying that they are analyzing this in any case. So, this paper does not suggest a single zoonotic event, although this seems to be lab leak opinion as you are well aware. Did you know that Bloom's next-strain analysis of Lv also has this disclaimer These results suggest that phylogenetic evidence alone is unlikely to identify the origin of the SARS-CoV-2 virus and we caution against strong inferences regarding the early spread of the virus based solely on such evidence."

@DanielNadolny Your contention is that it happened in Wuhan and DEFUSE said FCS insertion at the location it occurred in SARS2 (contradicted by the fact that it happened at S1/S2 and not at S2' and DEFUSE proposing S2' reflects subject knowledge of natural occurrences). Besides the fact that recombination explains the backbone of SARS2, insertions are known to cluster at that particular location and the more likely explanation is that FCS was likely picked up naturally in an intermediate host. This could explain all of the anomalies of the FCS that would be low probability choices for an engineer but are high probability for natural origins.

https://virological.org/t/putative-host-origins-of-rna-insertions-in-sars-cov-2-genomes/761

@DanielNadolny

Regarding the paper from Lv et al, yes I am very familiar with it. Its argument for A0 being ancestral is exactly the same as Jesse Bloom's argument in this paper -- https://academic.oup.com/mbe/article/38/12/5211/6353034 -- and Lv et al did not identify any new supporting evidence or conduct any analysis that provides quantitative support for A0 as the ancestral genotype. It's entirely a qualitative argument that goes like that:

1. There are genomes 1 mutation more similar to RaTG13 et al than SARS-CoV-2 lineage A (which is 2 mutations more similar than lineage B)

2. There was only one spillover to humans

3. The single genome that spilled over was certainly one of the ones one mutation closer to RaTG13

4. The genome with the mutation C29095T (or U if you prefer; this is "A0") was the earliest one of these that was sampled, so it is the most likely to be the earliest genome

If you're familiar with the "proCoV2" papers, the argument is the same until you get to step 4. Those authors prefer C18060T because their analysis says so, because their analysis basically counts how often a mutation happens and C18060T was very frequent early in the pandemic in the United States (e.g. early Washington state outbreak).

Earlier papers than any of these had previously identified C18060T and C29095T for the same reasons. I have never seen a quantitative analysis that takes sampling dates into consideration that supports either as being part of the proximal ancestral genome. Ironically, an analysis allowing multiple spillovers rather than a single spillover would make it more likely that either genotype is ancestral.

FYI and I don't think this has been reported anywhere else, the samples in Lv et al clearly overlap with a samples from a previous paper from Shanghai. It's hard to match samples perfectly, but it's very likely that the "A0" sequences are counted twice in that paper. Once for samples sequenced and reported in the earlier paper and once again for different samples from the same patients for this paper. I do not think that this effects any conclusions in the paper, because, like I said, the conclusion that "A0" is ancestral is a qualitative argument that does not depend on any data in the paper.

@DanielNadolny

I'm not sure why you're highlighting the first time the idea appeared in the media as some kind of evidence for or against a lab leak.

My point is to suggest a thought experiment: place yourself in early January 2020 and imagine you could pause time and study all of this for yourself (all of the relevant scientific literature about research in Wuhan and so on). Although the source in that RFA article is very much full of shit (check out the other articles they are quoted in), I think it'd be reasonable to find both lab and market origins plausible at that time. After all, Zheng-Li Shi herself did. A flat prior seems reasonable enough when no one had any clue what a retrospective case search would find or what the genomic data would show [1].

So in early January 2020 you know: (1) SARS-like virus, (2) most early patients linked to market, (3) rumors market sold wildlife on social media

Try to imagine learning about the evidence as it came in, from that point in time, and using Bayesian inference.

[1] The genomic data is important because no one can cover up the sequences of genomes that were exported from China; both the genomic data and the trajectory of the numbers of cases identified in travelers constrain possible scenarios in Wuhan even if you think there's an infinite capacity to cover up anything within China. If you throw out all the data from China, all of the conclusions you can make will be similar, except with wider confidence intervals.

@zcoli Here's an update to what the consensus view of lab leak was coming from the USA in on 20 February 2020 from VOA Chinese: https://www.voachinese.com/a/expert-says-coronavirus-may-comes-from-lab-accident-20200219/5295356.html

It describes a manuscript that was briefly on ResearchGate and then retracted by the authors. Of course, this led to suppression narratives that make it more suspicious. The whole manuscript is back online here if you want to read it: https://img-prod.tgcom24.mediaset.it/images/2020/02/16/114720192-5eb8307f-017c-4075-a697-348628da0204.pdf

The paper said the Wuhan CDC laboratory had a batch of animals for testing, including 155 bats caught in Hubei and 450 bats caught in Zhejiang. A bat attacked a researcher, during which the bat's blood came into contact with the researcher's skin. He therefore self-isolated for 14 days. In another incident, the bat's urine came into contact with him, so he self-isolated again.

The paper states, “The patient’s genome sequence is 96% or 89% homologous to the bat coronavirus ZC45. This coronavirus was originally discovered in the horseshoe bat, which is not native to Wuhan but lives about 1,000 kilometers away. But this bat is exactly what the Wuhan CDC used before the outbreak.

Sounds familiar, right? How could a virus get from 1000 km away in Zhoushan island to Wuhan? It must've been the researchers working there! But this is 1000 km in a different direction from Wuhan towards Shanghai rather than Yunnan province.

It's also an interpretation of a paper that was simply wrong, which is presumably why the authors retracted their manuscript. The paper with 155 bats in Hubei and 450 bats in Zhejiang is here: https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1003159

The samples didn't go from Zhejiang to Wuhan CDC. They went from both Hubei and Zhejiang to a lab in Beijing, where they were sequenced. You can see that here, for example, for one of the viruses sampled from a bat in Huangpi, a district of Wuhan city has a submitting lab in Beijing: https://www.ncbi.nlm.nih.gov/nuccore/JX473273

That's the difference between the "lab leak" theory and the "zoonosis" theory and it's why I posted the Weibo post from 30-Dec-2019. The "zoonosis" theory has been essentially unchanged. The earliest lab leak theories are hardly recognizable because they were repeated rejected and replaced with new theories, with people only becoming more convinced of lab leak along the way.