>They also thought they were going to get away with it. “The truth is never going to come out ”, wrote Rambaut in one message.
This is really mischaracterizing what he's saying in that message. He wasn't celebrating getting away with a coverup, as you're saying, he was saying that IF it was a lab leak, absent concrete evidence (which was very unlikely to appear), that we'd never know that to be conclusively true.
And broadly, because a lab leak is necessarily accusatory, the evidentiary threshold is higher than the baseline, normal, zoonotical origin option. And I don't think that's really unreasonable?
I really don't have a dog in this hunt; I think both options are entirely plausible. But you really seem to be overplaying your hand on this, and the above sort of mischaracterization isn't something that I'd normally expect to see from you.
> because a lab leak is necessarily accusatory, the evidentiary threshold is higher than the baseline, normal, zoonotical origin option
Chris, I strongly disagree with this statement. In medicine and other sciences, the “baseline” is that the treatment or hypothesis one is proposing is false and evidence needs to be gathered to “reject the null.” This is done because most treatments or hypotheses tend to be later disproved.
Here there are no such baselines. It’s not as if in the past few years, thousands of deadly coronaviruses were discovered, and say 99% were natural and 1% were developed in the lab. This was a pretty novel situation.
And words like “accusatory” belong to criminal trials or politics, not scientific research.
There's no way for them to say "we think it was a lab leak" without being accusatory, though.
I get that this is contentious, and fair enough. I would have liked them to be a little more "we'll probably never be able to figure it out", personally. But I think there really is a baseline here, given the origins of MERS and SARS, and that that making an accusation -- even a weak, caveated one -- in the absence of any evidence at all, is really tough.
Mostly my beef is with the way that Nate's characterizing this though. It feels way overconfident, and bad-faith in a way I really really don't associate with Nate.
I don't think Andersen et Alia did that. Consider what they wrote in their conclusion:
"Although the evidence shows that SARS-CoV-2 is not a purposefully manipulated virus, it is currently impossible to prove or disprove the other theories of its origin described here. However, since we observed all notable SARS-CoV-2 features, including the optimized RBD and polybasic cleavage site, in related coronaviruses in nature, we do not believe that any type of laboratory-based scenario is plausible.
More scientific data could swing the balance of evidence to favor one hypothesis over another."
That does not sound at all like "lab leak is a conspiracy theory." In fact, that sounds a lot like "current data is insufficient to rule out all hypotheses." Though, they did say current data was sufficient to rule out an engineered virus.
It is unavoidably accusatory to say "we think it was a lab leak". But I think you're espousing a sort of genteel politeness ethic that says we never point the finger without a disproportionately large amount of evidence. This is dangerous, because it systematically tips us towards natural origins, and in general, if it were a norm, avoiding questions about accountability and culpability in any question or field at all.
You can ask questions about someone's culpability, implicit or not, without making direct accusations. You do run the risk of being sued, like the Data Colada folks, but somehow I think we need to find a way past that reality because it seems toxic and protects powerful bad actors.
I think it is naïve to just consider this issue in terms of good science. This was a geopolitical issue. "The science" would have had geopolitical effects. Publishing something that has no evidence to support it but will have geopolitical effects seems to demand a bit more of a cautionary approach.
Chris C has a very good point, but he goes further than he needs to. It does not matter if it is or is not accusatory. All that matters is whether China's government, the CCP, would have taken offense. I think the clear answer is "yes". When Australia called for an independent investigation into the origins of covid, the CCP blacklisted Australia. It is clear that China would have taken offense because they took offense for less. What would have China done in response? Quite possibly exactly what they did to Australia, blacklist any country that speaks up questioning the origin of covid. China accounted for upwards of 50% of PPE production at the height of the pandemic. If they had blacklisted the world, or the west, that may have actually increased the severity of the pandemic, i.e. more people, in particular medical staff, may have contracted covid and died. So, at the very least, there is a chance that publishing the uncertainty about a lab leak theory or even going on record and saying it is quite plausible could have lead to more deaths.
But, if China did blacklist a bunch of countries after some publication of the plausibility of a lab leak, what would politicians in those blacklisted countries done? Consider here in the US, we had Donnie Trump. Is it plausible that Donnie would have interpreted China clamming up when accused of leaking covid in a lab, accidentally or otherwise, that it is worse than an accidental lab leak? Maybe, a bioengineered weapon that got out and killed 1 million US citizens? The US has gone to war for far less.
So, a potential consequence of publishing the plausibility of a lab leak is basically world war 3. I think this is not the most likely outcome, but it is plausible enough that it needed to be considered.
What immediate benefits would we have realized by pushing a lab leak theory? None. Would learning that it came from an accidental lab leak helped us develop vaccines quicker? No. Would it have allowed us to find drugs that could make the disease less severe? No... well, possibly; maybe the lab would have had special knowledge. Given that they shared the genome with the world, that's likely not true. So, letting the lab leak theory out had 0 immediate benefits. The long term benefits, of course, are the integrity of science, in particular of public health, virologists, and medical researchers.
So, ultimately, I think a cost/benefit analysis based on likely geopolitical consequences of letting the lab leak theory proliferate, is squarely in favor of tabling the origins question until the pandemic was over. Scientists do not live in an ivory tower. When their uncertainty, if it got out, can only have negative consequences, table the uncertainty.
This argument is really condescending and self serving. Everyone outside the pharma industry sees the obvious need for accuracy, transparency, and honesty on these questions.
Also, if "considering the geopolitical consequences" is going to be standard for how scientific investigations are conducted, how can we ever trust science?
Not a single person in this debate is associated with the pharma industry. The authors of the paper we are discussing are all researchers at either universities or non-profit research organizations, Scripps in this case. I'm not associated with the pharma industry. Everyone in this debate sees accuracy and honesty as valuable.
Transparency is another question. About How much should I be transparent? I'm writing a paper on a specific statistical estimator. I've got hundreds of pages of math derivations, thousands of simulations, entries in my research journal recording my suspicions about what the results should look like, debates with colleagues. All of that needs to be shared? Nonsense.
What did they, in fact, hide from us? Was it evidence? No. Was it theory that could help us understand? No. They withheld their personal opinion that a lab leak was plausible; in other words, they did not share their priors with us. Was it true that, at the time, the totality of evidence supported a natural origins explanation? Yes. Was it true that there was no solid evidence that supported a lab leak explanation? Yes. The evidence they had that supported a lab leak was circumstantial, the Wuhan lab is right there, or did not distinguish between the two hypothesis, the molecular evidence was consistent with both explanations. But, they should have shared their priors with us? Why?
So, in this case, you are claiming that researchers should be transparent about their unsupported hunches. Hunches that could have significant geopolitical effects if politicians understand them incorrectly, given who the US president was at the time that was a very likely possibility. Given that everyone and their mother was citing papers and drawing drastically wrong conclusions from those papers, it also seems premature to publish unsupported hunches about a lab leak.
This issue has nothing to do with "how scientific investigations are conducted." This is about what should be published. Should we publish unsupported hunches that could cause a geopolitical storm? No.
The research into COVID was not standard. There is no reason to suppose that what we should have done with regard to a virus that was killing millions should be standard operating procedure. An analogy: consider battlefield medicine. The gold standard protocols of hospitals are likely not going to be adhered to when treating a gun shot wound on the battlefield. Does that mean we should adopt the battlefield medics procedures for our hospitals? Likely not.
Why publish what was NOT their best considered view? Acting on any one of Nate's possible motivations is wrong, damaging to science and the public' ability to gain information from scientists.
They did publish their best considered view. In particular, their position, as published in 2020, was that the lab leak hypothesis had not been ruled out but the totality of evidence made it implausible. What they did not publish was their suspicions that the lab leak explanation was more plausible than the data implied.
Scientific journals are not in the business of publishing scientists' unsupported, a prior hunches. They are in the business of publishing evidence and the conclusions that follow from that evidence. Journals leave room for scientists to speculate on further areas of research wherein unsupported hypotheses may be published. But, this is not a requirement.
Just looking at this now. It’s important to remember that Trump told the American people again and again that Covid was not dangerous, like a common cold, we would all be back in church by Easter etc. the origin issue was not important. Those were the comments that were remembered in November.
There is a difference between "concealed" and "not published".
I can think of many reasons why it should not be published.
1) It will do nothing for anyone trying to understand the actual results that are supported by evidence.
2) There is so much of it that publishing it all will likely just lead to wasted paper and ink if it's printed or electricity and storage space if kept digitally. Note there are millions of scientific papers published each year. If each one of those papers required publication of all research notes, even more would go unread and thus be wasted.
3) No one will read it, especially if every paper published was also published with all the work that went into it.
4) It would take me hours and hours to collate it into something that would make sense to someone else. If every research did this, it would amount to millions of researcher-hours wasted each year.
4) There is likely to be claims amongst my research notes that are not justified by data but they are my hunches.
5) Someone might cite my unjustified hunches as the testimony of "an expert". (Notice this has happened several times over the past 5 years. Also notice, politicians and those with agendas have used the research notes and debates amongst researchers to push policies and narratives that are not justified and also to undermine science as an institution).
7) Does someone have to peer review the notes? If so, that amounts to even more researcher-hours wasted.
"I think it is naïve to just consider this issue in terms of good science"
"The long term benefits, of course, are the integrity of science"
You can argue one or the other, but that a series of geopolitical calculations for what goes on in a science journal isn't going to lead to the second thing.
I can argue one or the other of what? One of the two quotations? So, either we only consider it from good science or also consider more?
Are you saying that doing all that geopolitical calculations about what to publish is not going to lead to the long term integrity of science, the 2nd thing?
If so, well duh! That was my whole point. You have a choice: 1) maintain the integrity of science in the public's eye but possibly lead to many more deaths, and worst case a war or 2) compromise on the integrity of science but maintain relatively stable geopolitics.
Anyone that chooses (1) has twisted priorities, in my opinion.
Why couldn’t Andersen et al have published a full record of what information was available, in order to either use it to produce a probabilistic conclusion about the origin of the virus, or use it to explain why the evidence wasn’t sufficient for such a conclusion?
It seems pretty far-fetched that such a paper would cause China to ignore a gigantic PPE sales opportunity, let alone cause Trump (who was, for all his faults, relatively against sending troops anywhere) to basically end human civilization...
It seems Pretty far fetched for China to do things that it has actually done? When Japan detained a Chinese fishing trawler for violating it's maritime boarders, China shut down exports of rare earth metals to Japan, despite this being a billion dollar revenue source for Chinese miners. When Australia called for an independent investigation into the origins of covid, China proceeded along a line of first import restrictions of Australian goods and then export restrictions of Chinese goods to Australia. When a democracy activist in China received the Nobel peace prize, China reacted by shutting down imports of Norwegian fish. it is not far fetched to suppose China will do what it has been doing anytime some country offends it.
You seem to think they withheld data and evidence. They did not. They merely withheld their opinion that a lab leak, despite having no evidence to support it at the time, was still plausible.
Y'know what, while I do disagree with your wider calculus on this (so I guess I'm twisted), yes I misread you on the second sentence. So my bad, got the wording wrong.
Also I don't quite see how these particular papers had to reach the scale of the conclusion on any of this stuff in the first place. It's not like they *had* to put down it was leak or not leak. But as it played out, seems like this became an easy citation to act as if that part of the debate was finished
It is already the case that Institutional Review Boards, IRBs, consider the potential harms and benefits of research. If the potential harms outweigh the potential benefits, IRBs will not approve a study. Thus, researchers have to consider extra-scientific considerations, in particular ethical and legal considerations.
Geopolitical risks are potential harms. Thus, IRBs will consider that. Since the researcher wants to perform their study and get it published, they have to consider potential harms that an IRB will consider. Thus, scientists will be obliged to consider the CCP's reaction when their study has geopolitical consequences.
Just because the vast majority of scientific research does not have such stark geopolitical consequences does not mean such geopolitical consequences can be ignored.
I feel there are avenues for you to object here. You could argue that the political reactions are special and IRBs are not in a good position to judge such geopolitical risks; this avenue doesn't really settle the question, though. You could argue that the geopolitical risks I outlined are very unlikely to occur; thus, IRBs can ignore them. Or, perhaps, you have another avenue of argument.
Maybe, you can actually share an argument instead of making a dogmatic statement.
I don't for one second believe that the Scientists publishing what they know not to be true was to prevent WWIII. I think the motive that Nate ascribed are more likely than what you postulate. Publishing scientific flim-flam is worse than publishing nothing at all. And for the MSM to write the findings in stone is why they are disappearing faster than polar bears.
"Scientists publishing what they know not to be true" none of the researchers that are a subject of this discussion are guilty of this.
Rather, Andersen et alia stated that the evidence cannot rule out a lab leak explanation but that the totality of evidence available in 2020 made it implausible. This was true. It was not false. They knew it to be true. Hence why they published it.
What has come to light is that Andersen et alia thought the lab leak explanation was more plausible than the data available entailed. This is not a statement of knowledge. It is a statement of personal opinion. It is a suspicion that the data is incomplete and when more complete data is made available the more complete data will show the lab leak explanation is more plausible. They chose not to publish this statement of personal opinion that was not supported by any data available in 2020. I do not see anything wrong with refusing to publish an unsupported personal opinion.
There was a huge issue, in early 2020, in getting the world to take COVID seriously. How much more seriously do you think people would have taken it if, on the first of February, global headlines had read "leading scientists believe COVID probably escaped from a Chinese laboratory", as was in fact their belief? How much more quickly and thoroughly would travel and other restrictions have been put in place? What effect would this have had on the spread of the disease? Ultimately, how many life years would have been saved? I think the answer to the last question is in the tens to hundreds of millions.
If in doubt, just tell the truth. Particularly if that's what you're paid to do. And you have no sane reason to believe that you're an expert on global geopolitics.
We've allowed the incentive structure for senior scientists to become seriously misaligned. I suppose one way of realigning it would be to follow the example of the British Navy in the 18th century - public executions of everyone involved would encourager les autres to take a different view of their self interest next time around.
I honestly don't think that would have made any difference because Covid is contagious enough (and contagious while still asymptomatic) that absent a full 100% travel blockade, it was going to make its way across every border and then spread internally regardless.
There were a couple island nations that (mostly, at significant cost) made this work; it just wasn't an option for most of the world. It was already in mainland Europe in January 2020, at that point it was going to make it just about everywhere.
It's clear it would have made some difference - it was a significant part of the success of Japan, New Zealand, Australia, Taiwan, etc. Islands, but with hundreds of millions of people. A complete quarantine of China on the first of February would have changed everything. There are valid questions about how much difference it would have made, but claiming it would have made no difference at all is silly.
The problem with tabling the discussion (and I am of the opinion both are highly plausible and we are unlikely to ever know for certain), is that tabling it essentially renders any hope of clarity completely absent, prompt evidence collection is the only hope of ever plausibly investigating the origins, and the outcome of that knowledge is far greater than simply “scientific integrity”.
It matters what the answer is because if it was inadvertently released from a Chinese lab that has huge implications moving forward for when and how we scientifically collaborate, and the world in aggregate was so deeply harmed by their casual laboratory safety measures, the world is capable of pushing on China to clean up their house. It also brings into sharp clarity the danger of all gain-of-function research and would propel increased oversight and needed regulation of these activities. Some of this may happen anyway, but the answer to the question is not idle. And given that scientists are, even now, doing things (which we know from published papers) like putting Omicron spike proteins onto Delta Strain to see if it gets worse (it does) and are carrying on this research in multiple labs around the US, and more we probably aren’t aware of, the theoretical gain from identifying the origin and recognizing risks of lab leaks from such research is all around us today and will continue to be present. This isn’t an abstract future benefit but a plausible near term benefit. We know of other instances in the last three years of COVID escaping labs where scientists have been studying it, lab leaks are more common than you might realize.
And honestly, from the Chinese perspective, they aren’t much happier about their wet markets being blamed for the origin than their lab. It might seem to you as if they’d strongly prefer one, but they don’t actually like either one of them, and have obfuscated evidence for both.
No one is calling for tabling the question indefinitely. Rather, table the question until the pandemic is handled. The question is valuable, in the long term, for exactly the reasons you mentioned in your second paragraph. But notice, all those reasons are long term. None of them would have helped save lives in 2020.
You are right. Tabling the question until later may cause data collection problems. But, I'm less worried about that. We have genetic structure of the early COVID variants. China, presumably, has as good a record as possible of the the spread in Wuhan. There is not much more you can use that plausibly would have been available back in 2020.
Your last paragraph has merit but is also wrong. Sure, they don't like either. But, they like the lab leak explanation far less than the natural origins in the wet market explanation. They blacklisted Australia because Australia basically suggested that a lab leak explanation was plausible. They did not blacklist anyone for pushing a natural origins at the wet market explanation.
There is a sort of baseline on recent pandemics from straight zoonosis (several) and from lab leaks (1977 H1N1). So there are priors favoring Z. But the aren't infinite. Even at 100/1 priors it looks like location, DEFUSE, the sequence data, probable discovery of ancestor in Vero cell cultures, near absence of lineage A from HSM, still no trace of intermediate wildlife host,... will swing the posterior odds to favor L.
It's probably best if governments keep mumbling about inconclusive results, but meanwhile we need international regulation of GOF, especially in places like Boston and the Netherlands, where some really crazy shit has been done.
> And broadly, because a lab leak is necessarily accusatory, the evidentiary threshold is higher than the baseline, normal, zoonotical origin option. And I don't think that's really unreasonable?
That could be a reasonable answer if Proximal Origins had withheld judgment and said that the evidence is inconclusive. But that's not what the scientists said: they told us we *could* draw a conclusion, and that the evidence likely ruled out a lab leak.
And the paper did lead to accusations, just not against China, but rather against the American "racists" and "conspiracy theorists" who defended the plausibility of a lab leak. If the organizing principle was to avoid accusation, why didn't scientists stand up and defend people against these attacks, given that they privately believed a lab leak was plausible?
I think it reads like someone who privately understands what the others are saying about strange “coincidences” and such, but also understands just how bad it would be for his field if the lab leak hypothesis were to gain traction, so is trying to shape the conversation toward dismissal by pointing out there will never be conclusive proof of a lab leak -- he’s saying it’s a safer bet to exaggerate their confidence in a zoonotic origin.
I would characterize myself as strongly favoring the lab leak hypothesis and strongly believing there was a cover-up. I clicked through and read the source, and came to the same conclusion as Chris -- Nate badly mischaracterized this quote. The scientists are not saying what he accuses them of. He should correct this.
I'm sad to say I have seen Nate take highly implausible (to me) readings of sources before. At this point, I don't especially regard "reading text and generating unbiased analysis of what the people generating it meant" high among his skills. (I think that's a more likely explanation than that he is being deliberately unfair or biased).
Where did Nate state or imply that the authors were "celebrating" something? Their mindset was that the deception would stand, and "The truth is never going to come out" was part of that mindset.
They're acknowledging what they think the truth is and that there will be no conclusive evidence, therefore opting for the safer, more opportunistic, career-oriented approach. Tbh I'm not sure why they're getting defended so hard here.
If you're going to accuse someone of mischaracterization, don't mischaracterize what they have said. Nate didn't say that Rambaut was celebrating getting away with the cover-up. He said that Rambaut and others thought they would get away with the cover-up. He backed this up with a quote from Rambaut which seems to me to provide evidence for that assertion.
You, on the other hand, don't have a quote providing evidence for your assertion that Nate was saying that Rambaut was celebrating the cover-up. That's because Nate didn't say it.
I would normally think that your misunderstanding might be linked to you already having strong preconceived views on this subject, which have lead to you searching for a way to discount the damning evidence Nate sets out. But reading on, I see that you say that you don't in fact hold such views, so I must be mistaken.
I directly quoted Nate, and obviously the whole post is up there for people to read. I think it's very clear that Nate's attributing that specific quote, "the truth is never going to come out", to them "thinking they could get away with it", meaning Nate's asserted cover-up.
But read the actual context of the quote; that's not at all what Rambaut is talking about.
Rambaut is talking about the truth not coming out. Which, on the plain reading of what Nate wrote, is what Nate is saying Rambaut was talking about. You read something into what Nate wrote that wasn't there.
Lets give a full quote of what Nate says since that is in dispute:
> They also thought they were going to get away with it. “The truth is never going to come out ”, wrote Rambaut in one message.
That is very clearly accusing Rambaut of being dishonest here, of thinking that lab-leak was plausible and much more likely than Rambaut had published. The rest of the article is accusing them of that too.
But had Nate included the very next words from Rambaut in that message "(if escape is the truth)" he would've, and we would've immediately seen that the context absolves Rambaut of that accusation at least in that message. It's a discussion of a hypothetical of how *in the world in which COVID started as a lab leak* there wouldn't be evidence available to prove it (due to particulars of that situation, of China's secrecy, whatever).
There really isn't a valid argument to the contrary, Nate was misleading on presenting that quote from Rambaut and should be correcting his article.
No. He accuses Rambaut and his co-authors of being less than honest - but not here. He accuses them of being less than honest with reference to the private messages that show that their public statements misrepresented their actual beliefs.
Here, he states that Rambaut thought that the truth would not come out. Which is only suggestive in combination with the other evidence. By itself, it means nothing.
Anyone would think that those fixating on a simple, accurate, and appropriately sourced statement of relatively minor importance were doing so because they desired an excuse not to have to think about what is staring them in the face.
Nate didn't say they were "celebrating" a cover up. He said they thought they would get away with publishing a purposefully misleading article. Clearly, they were confident in the deception, although KA worried about what the spooks might know.
Also, accepting lab leak would result in stronger safety policies. For that, the evidentiary threshold should not be so high.
No. He’s saying lab leak is plausible. But he feels he can lie about that because he will never be proven wrong definitively.
And taking your two assertions together (higher evidence threshold & high unlikelihood of concrete evidence) it would be nearly impossible to suggest a lab leak leak had occurred under any circumstance.
I really do think that there was a hard turn after Trump was elected by center-left media outlets to fight fire with fire, meaning if right wing media outlets were going to lie in order to push an agenda, then they are going to do it to. That's not to say it never happened, but Trump getting elected made it even more prominent. In this instance, because 'Trust The Science' was the motto being pushed, anything that might undermine that message wasn't reported on. Needless to say I think this is an extremely bad direction for mainstream media outlets to go.
The old line (think Bush era) used to be that the left lacked a Fox News, and was kind of boxed in by journalistic norms whereas Hannity could do what he does every night. There was a point there on several levels - however, the internet changing the nature of media platforms and a reliable college-educated audience has made that something of the past. And yeah, creates some *very* bad incentives.
It was, and is, all about Trump. Anything deemed helpful to him, in any way, must be discredited, sneeringly dismissed and suppressed. Anything that is thought to hurt him must be propped up, repeated, treated as gospel truth - no matter how shaky it really is.
It feels like the left, having been fairly defensibly the "party of science" for the past few decades, has learned the lesson that they don't actually have to have beliefs that align with science or even do science at all - they can just skip to the end, proclaim loudly that "the Science has spoken", and that any other position is a conspiracy theory not worthy of a moment's attention.
With Biden admin (finally) cutting funding to WIV, it seems the slow walk from "lab leak is a racist conspiracy theory" to "well, shoot, we don't know what happened" to "it was a lab leak but it would be impolite to say so" is well underway.
Biden himself always seemed like a quiet lab leak believer. I think one thing that really pushed scientists against the lab leak theory is that gain of function research was very contentious and it wasn't until 2017 when it was allowed in the US. So to have a possibility that gain of function research had the worst result possible only 3 years into it would lead a lot of scientists to really want to hammer home the zoo tonic origin and ignore anything that might point in the other direction. Hell, it seems pretty clear that covid wasn't a modified virus, but even the optics of where it escaped from would put gain of function research in the crossfire
One thing that some in the media did was try to conflate the "it's bio terrorism from China" theory with the lab leak theory. It was pretty unlikely and remains unlikely that this was an intentionally bioengineered pandemic, but early on the lab leak theory was treated the same way.
Kristian Andersen has stated many times publicly that (1) he initially believed that a lab leak was likely, (2) he subsequently changed that view based on the scientific considerations that are detailed in his scientific publications, and (3) he has always believed, and has always written, including in the Proximal Origins paper, that a lab-leak scenario is possible (though he now doesn't believe it's plausible). He explained this in detail in an all-day interview, half of which was conducted by a House Republican staffer, as well as in his prepared testimony:
Nothing in Nate's post or in the links contradicts Dr. Andersen's version of events. In fact, none of them of even attempt to. Nate quotes Dr. Andersen's February 1, 2020 email as if it were a smoking gun that Dr. Andersen was lying, without even addressing the fact that Dr. Andersen has repeatedly stated that he changed his mind between February 1, 2020 and the time of publication.
I mean it's obvious that Dr. Andersen did change his mind at some point! He has been publishing papers to the effect that the lab leak is unlikely continuously since early 2020. What is the argument here? That his initial impressions from February 1, 2020 were his true and final feelings and that everything he has said and published since then is disingenuous? That's absurd.
To be clear, Dr. Andersen could well be wrong. I take no position on how plausible the lab-leak hypothesis is, because I don't have the training necessary to evaluate the scientific evidence. Plenty of people who do have the training take different views of the evidence and that's fine.
But I do have the training to evaluate allegations of fraud. And I see no persuasive evidence of it here. Until I see a careful, good-faith assessment of the record that considers Dr. Andersen's detailed explanations and attempts to refute them, I consider the allegations of scandal completely unproven. And this post, which asserts that the facts are "simple and clear" without even engaging with Dr. Andersen's public statements on the issue, and mischaracterizes the statements that it does cite (see Chris C's comment) gives me zero confidence that Nate has made a good-faith effort to get to the truth.
> he has always believed, and has always written, including in the Proximal Origins paper, that a lab-leak scenario is possible (though he now doesn't believe it's plausible)
In April the Slack shows Andersen privately worried that he can't disprove a lab leak. He even posits a potential mechanism by which it could be engineered:
> [Andersen]: We also can't fully rule out engineering (for basic research) - yes, no obvious signs of engineering anywhere, but that furin site could still have been inserted via gibson assembly (and clearly creating the reverse genetic system isn't hard - the Germans managed to do exactly that for SARS-CoV-2 in less than a month).
How does this statement square with Proximal Origins, which says "we do not believe that any type of laboratory-based scenario is plausible"? He just posited a scenario by which it could happen! This scenario does not appear to rely on any kind of fantastical black swan event -- quite the opposite, it says that the Germans did exactly that in less than a month. I don't know how to read this statement as anything other than "an engineered virus is plausible."
If this is going to come down to weasel words, where "implausible" can mean "I privately admit that it could well have happened that way," then they are entitled to their weasel words but they have lost my trust.
1. Implausible and "can't fully rule out" are consistent with each other, as you acknowledge. In fact, the March 2020 paper expressly said that lab leak hadn't disproved. You can call it weasel words, or you can call is "accurate use of the English language."
2. When all of someone's private communication on a subject have been subpoenaed and publicly released, someone attacking that person based on a single message has the burden of proof of showing that the document is being read fairly in context. There has been no effort to do that here; there is no mention at all of the context of the message or what came before or after. Before I can even begin to criticize Andersen, I'd like to see someone analyze the full record and show how it contradicts Andersen's public statements. That hasn't been done.
3. Nate and others are accusing Andersen of fraud. That is an extraordinary claim and nobody has come close to proving it or even seems particularly interested in doing so in a careful and fair-minded way. If, after all of someone's private communications have been publicly released, the best that you can do is identify one isolated message and assert that is shows the scientist used "weasel words," this strongly suggests the fraud accusations are baseless.
I agree that fraud is a strong accusation, and based on what I’ve seen so far I don’t know if I would go that far.
That said, the authors’ main defense is to hide in the ambiguity of what the word implausible means. When their colleagues are citing their paper as evidence that Covid is zoonotic, and signing letters condemning “conspiracy theories” that it could have been a lab leak, then “implausible” is taken to mean “we can dismiss doubters.” But when their own private doubt is made public, then “implausible” is taken to mean “we were always open about the fact that we hadn’t ruled it out completely.”
They can’t have it both ways. If the most expert scientist is allowed to have private doubts, then everyday people are entitled to have public doubts.
If scientists are resorting to this kind of argument then their entire enterprise is toast. The public will not accept arguments of the form "scientists speak a very special form of English in which nothing they say can be taken at face value regardless of how clear it may seem". Why fund people who make such arguments?
There's a lot of context about the messages available. Nate Silver is letting you know more messages exist, it's on you to go read Racket and Public if you want to see more of their logs. But be aware: it doesn't get less damning, it gets moreso.
They just lied. Yes, they engaged in a kind of fraud. If they were doing this sort of thing in the private sector and about accounting instead of deadly viruses they'd be fucked already and probably on the way to jail. Why is it so hard to accept that? It's called lying, not "accurate use of English". Academics are not morally purer than normal people, frankly the evidence is strong that they're dramatically more corrupt than normal people are. It's a culture that doesn't seem to have any norms of honest behavior. Perhaps unsurprising given its essentially communist setup.
Why demand that scientists, among themselves, contrive to speak so every possible raging conspiracist dumbass is pre-emptively neutralized? And no, you don't have the full context. You just want to believe you do.
Yes, the populist attack on science and education has blinded liberals to the truth about scientists. They are human beings like everyone else. They have their personal stakes, their preconceptions, their ambitions, and their blindnesses. This is not a new thing, it has always been thus. (If you doubt this, you ought to read Charles Mann's 1491 and follow the endless roadblocks scientists threw up along to the implications of archeological finds that shed light on the history human beings in the Americas.)
Scientific inquiry is an incredibly powerful way to seek the truth of many matters. It has shed light on myriad things, and will continue to do so. But it is always a work in progress, and can never be personified as the final word. To do so is to treat science as your religion -- which is exactly what many people do. Further scientific inquiry will upend a fair portion of what appears true now, and it will update and adjust a great deal more. And that would be true, even if individual scientists didn't have all the personal weaknesses shared by the rest of our species.
As a physician who has followed your work across domains for about 15 years I’m astounded by the how wrong have gotten this Nate. Sad to see your name behind this weak, conspiratorial to be point of being delusional “reporting”. 538 is worse off without you but you seem to be missing your edge now too. Unsubscribing but maybe will circle back in a year or two.
It would be great for you to break down why Nate is wrong instead of simply asserting it. The lack of public back and forth on this is extremely frustrating.
That's a little bit out of date, I've found more arguments in favor of natural origin since, but it should give you an overview, and you can dig into the linked sources if you need more.
I also agree that it would be nice to have more public debate on this. I'm actually in negotiations with Rootclaim to debate this for cash, though it's still TBD whether we'll agree on favorable terms and when the debate will happen.
Wow, you seriously think, with all the data about ) gain of function research being done in Wuhan, 2) the fact that years later, no animal source for this leak has been found (as compared to other, truly natural outbreaks), 3) the ONI report, and 4) even comments from the scam letter signatories that it looked engineered, that it's a natural outbreak? Astounding.
1) Do you have any idea how many places do GOF? Without that, it's impossible to evaluate the importance of this point.
2) Animal sources are almost never found viruses. We've been looking for an origin for MERS for a decade. This "Middle East" RS appears (possibly) to have come from bats in the vicinity of South Africa, thousands of miles away, made a leap to camels, and then to humans.
3) The report says nothing about origins, and doesn't present strong evidence for a lab origin. For example: "We continue to have no indication that the WIV’s pre-pandemic research holdings included SARSCoV-2 or a close progenitor, nor any direct evidence that a specific research-related incident occurred involving WIV personnel before the pandemic that could have caused the COVID pandemic.'
If every city had a GOF lab with DEFUSE-like plans for coronaviruses, that would certainly change one of my Bayes update factors, just as you say. It's highly unlikely that there are more than several, if any, others. There are no reports of repeated international warnings abut sloppy labs other than in Wuhan. Meanwhile, there are >20k wet markets, with only about 4 in Wuhan, so my simple population-based x100 based on Wuhan location is probably an underestimate.
1 is completely irrelevant. If the outbreak started in Wuhan then all that matters is what mechanisms were available in that location that could have caused it.
2 is also completely irrelevant if substantially more time/money has been spent in this case vs other cases. MERS was not a pandemic so it would hardly be surprising that less has been spent looking for it’s origins. Of course you’re free to provide evidence to back up your claim.
Re your point 2- the proximal hosts (camels and either civets or raccoon dogs) of MERS and SARS CoV1 were found. The proximal host of SARS CoV2 has probably also been found in sequences from Vero cell cultures sequenced early at the Sangon lab, with 3 SNPs differing from original SC2, all in the direction of the ancestral sequences.
It's difficult to overstate the damage done over the past three years to "trust the science" and "trust the experts" by these people. Before COVID I was firmly in the camp that thought the scientific consensus was usually correct, or at the very least had some merit. The climate change debate had weakened that trust somewhat already, but a lot of the problems there I had blamed on an overzealous media addicted to sensationalism and pushing an agenda. Now? That willingness to give experts the benefit of the doubt is gone.
Yes the real mind-bender is when you start to update past beliefs in light of this "new" (not really new) information. Climate stuff is not different to COVID it seems. The ClimateGate emails were that field's lab leak coverup equivalent, but because our society worships academia it was all swept under the carpet and nothing was done even when they were writing the most damning things.
I am utterly unshocked that you think the 'ClimateGate' emails were evidence of truth about climate being covered up. It wasn't swept under the carpet, it was investigated at tedious length. You're just mad because no evidence of such a coverup was found. You probably think 'trick' meant 'trick' the public', too.
One should not forget that Trump's correct statement that the Wuhan virus came from Wuhan was used as a stick to beat him by the media:
ABC:
“Trump's 'Chinese Virus' tweet helped lead to rise in racist anti-Asian violence”
Mar 18, 2021 — "Anti-Asian sentiment depicted in the tweets containing the term 'Chinese Virus' likely perpetuated racist attitudes and parallels the anti- ...
UCSF:
Trump's 'Chinese Virus' Tweet Linked to Rise of Anti-Asian Violence
Mar 19, 2021 — A group of researchers found Twitter users published more anti-Asian hashtags the week after he first tweeted about the “Chinese virus.”
As a scientist myself I'm kind of of two minds here. I think one major issue is that most of us are used to writing for other scientists to read, and so because we're all humans, I think a lot of scientists are trained to sort of correct for those sorts of biases that we know exist. That's not to say that the authors acted ethically--they obviously didn't--but I read a lot of papers that I think are crap but we just sort of say "I think that's crap" and then move on, or if we're really motivated, write a paper refuting it. We're trained very well for that, so writing a crappy paper isn't usually seen as a big scandal, it's just seen as sloppy work. What we're not trained for is talking to the general public. And for most of us, that's fine. In my situation, the chance of there ever being a global emergency relating to non-classical nanocrystal growth is rather low, so it's unlikely that there will aver be any sort of public policy or what have you based on my work. But it seems like for people in public-facing fields, that should be emphasized. I realize it can be difficult to determine sometimes what constitutes a public-facing field, but I feel like something like "public health" where "public" is in the name might fall into that category. So from that standpoint, even with as charitable as possible an explanation for their shoddy work, as public health professionals they need to understand that the general public is not trained to critically analyze science papers but is constantly primed to think about politics, so they need to be extra careful about controlling for political bias in their work if they know that the work is going to be of public interest.
I think there's also something to be said from the other direction as well--the media is absolute crap at reporting science. Before COVID that was sort of just something that was irritating but since COVID it's really exposed how much of a problem that is. The reality is that interpreting science is hard and those of us whose job it is to do that go to school for a very long time to be able to have the technical skills for that. Ultimately due to media biases they'll often just report on the things they already wanted to report on and pick scientific studies to cite that support what they already believe. The secret is that scientists do that too. We're all human and even though we try to control for our own biases, we do get excited about our own work and want to find references that agree with us. The hope is that if something bad was missed, that will get sorted out in peer review, and then if it doesn't the people who read the paper will read it critically. One thing I wonder about is that in the early days of COVID, a lot of journals instituted policies that would fast-track COVID-related papers for review, which was a policy that obviously made a ton of sense at the time but also likely led to overworked reviewers who likely had COVID-related research of their own and their lives, like everyone else's, were uprooted and all that, so I wonder if a lot of these reviews got kicked to postdocs and grad students who haven't reviewed a lot of papers before and let more slip through the cracks, and then what slipped through the cracks made it to the media, and then Pandora's box was open. As you pointed out, early COVID was a hard time for everyone, but I think that for those people whose job it was to provide and disseminate the best and most current information, including the public health scientists, the journal reviewers, and the media, there were a lot of failures that compounded and led to serious problems such as this one.
“the media is absolute crap at reporting science.”
It’s pretty safe to posit at this point that the media, with a few exceptions, is absolute crap at reporting almost anything.
Education polarization has made them overly trusting of certain sorts of authority figures, most especially academics/university administrators, non-profit workers, and civil service workers, and overly skeptical of others, most prominently police and businesspeople.
But that’s a small part of the problem which encompasses only a small fraction of stories and reporting.
The major driver is that “reporter” is no longer a job with the combination of pay and prestige, in most instances, to attract and retain top-notch talent. To be frank, most reporters I know or have read enough of to have some familiarity with are complete mid-wits with neither the raw intelligence, nor the work ethic, to actually analyze and critique almost *any* complex narrative in almost any field of endeavor.
And because most don’t actually have the ability to do this, the vast majority of reporting has become essentially editorial in nature, with a reporter’s meager “analysis” driven almost solely of their priors towards the actors involved.
Thanks for chiming in. It's good to hear from scientists and researchers when things like this happen.
Perhaps Nate went too far in characterizing the scientists thoughts on the Nature article. We don't know what people think. But Nature made sure to publish an article that loudly said COVID wasn't a lab leak. Frankly, scientists are loosing ground on trust from the public. Covidgate, p-hacking, faux scientific papers being published, before that was Climategate, etc. -- these instances feature peer reviewed literature and journals that are at least suspect if not part of the problem. We can waft some of these aside in some cases but in COVID's case it's Nature and Lancet, the premier publishers, so the problems with scientific literature gaining trust from the public has hit rock bottom.
Trust can't be gained from simply training scientists to speak more eloquently to the public. The scientific publishing process needs to divorce itself from whatever has tainted its well.
I support open science. But that would get rid of all the money that flows via the journals. Perhaps that money is the problem.
Wife and I were talking about that this morning, in this context and as it pertains to medical research with respect to nutrition.
I have no time or patience for physicians and scientists who bemoan the public’s lack of trust in science. Do they think we’re blind to the influence of money on what they call “science”? They work for institutions that whore them out to political and commercial interests in exchange for money and political credit, publish their work in journals that whore themselves out for money and political credit, grant their blessing to others’ shoddy work and attack work that challenges the mainstream narrative, again for money and political credit.
Having whored themselves out, they stand in front of the world and try to present themselves as “impartial scientists”? And bemoan our “lack of faith” or “lack of trust”? Faith and trust have no place in the scientific method, nor has the scientific community conducted themselves in a manner worthy of trust and faith in their collective integrity.
At the very least, Andersen, et.al. went into their investigation with a definite preference for the preferred outcome. They expended a great deal of effort looking for evidence that would support a plausible rationale for natural evolution, but did they work as hard to establish a plausible rationale for laboratory evolution? The content of their Slack channel suggests not. They wanted to prove one hypothesis, and disprove the other. I don’t believe they accomplished either goal.
The truthful conclusion of their investigation, such as it was, would have been “we simply can’t reach a definitive conclusion”. But that wasn’t the outcome they wanted, so they wrote the conclusion they and their patrons wanted and massaged their paper to support it. It wasn’t honest science, but it was very much in line with “science” as we see it conducted in the public sphere today.
Can I suggest that folks read the documents for themselves. Very clearly shows that they revised their paper to be more skeptical of lab leak after new data came to light in March. Prior to that, some were personally leaning towards lab leak without hard evidence for it, and being careful to distinguish publicly between what they reckon and what they can scientifically conclude. Very surprised to see Nate Silver swallowing the misleading framing.
Here's brief partial list of observed features that contribute substantial likelihood ratios to update the prior odds that favored Zoo. All the updates favor LL.
1. location: P(Wuhan|L, FCS) ~1, P(Wuhan|Z, FCS) <0.01 ratio >100
2. Intermediate hosts found: P(Vero cells but no wildlife|L)/P(Vero cells but no wildlife|Z) = big
3. FCS sequence: P(2xCGG|L)/P(2xCGG|Z) ~100
4. FCS neighborhood: P(3 synonymous SNPs|L)/P(3 synonymous SNPs|Z)= fairly big
5. restriction enzyme sites: P(few but no long stretches|L)/P(few but no long stretches|Z) >10
I'm going to take issue with all of your numbers. I presume Z=zoonotic origin; L=lab leak; FCS= furin cleavage site (?).
1) P(Wuhan|Z, FCS) is definitely not less that 0.001. It is also not the case that P(Wuhan|L, FCS)~1. The first stems from the fact that we know wet markets do make for solid breeding ground for virus that can jump species. The second stems from the fact that researchers and staff at the Wuhan lab interacted with people from all over China and the world. An asymptomatic research that infects his cousin who then travels back to Beijing is very plausible. Both of your numbers are questionable.
2) This point is irrelevant considering that COVID has been found amongst the genetic material of dogs recovered at the wet market. Thus, P(vero cells but no wildlife)=0.
4) This paper suggests that synonymous SNPs in viruses is a way for the virus to escape the hosts immune system. If true, that would entail there would be an adaptive advantage to having synonymous SNPs. Thus, synonymous SNPs would not be unexpected. https://pubmed.ncbi.nlm.nih.gov/31416386/
5) This one doesn't make any sense to me. Restriction enzyme sites are small sequences of code that bacterial enzymes bind to and cleave. Everything I've found say these are generally between 6-8 base pairs long. Thus, the probability that there would be few of them and they would be small would seem to be quite large for any virus, irrespective of origin. Thus, these numbers are suspect too.
In summary, all the numbers you shared are highly suspect.
Wow, a combination of misreading and misunderstanding conditional probabilities.
1. I used P(Wuhan|Z) of 0.01, not 0.001, although based on the distribution of wet markets one might justify the lower number. Your visitor story is just a reminder that P(Wuhan|Z) is not zero. But conditional on Z Wuhan is no more likely than other places with similar population, maybe less. Conditional on L, sure P(Wuhan|L)<1 because there are other labs working on coronaviruses, but this is a big one, known to be sloppy, known to plan FCS insertion. So P(Wuhan|L) is not a whole lat less than 1.
2. The raccoon dog story fell apart. See Jesse Bloom's reanalysis. SC2 RNA negatively correlated with stalls with potential hosts, other than humans. But it's true that on closer examination the samples with Vero cells and SC2 RNA are not a smoking gun, because that RNA could have come from humans rather than the culture cells.
3. The CGG numbers for this group of viruses are solid, ~0.03. For humans, ~0.36. There's nothing wrong with this calculation except that I was conservative in treating the 2 sites as independent under the L hypothesis.
4. Synonymous mutations are indeed common. They just seem way too common here, e.g. one stretch of 60 mutations, all synonymous.
5. Wow, the size of the restrictions sites is not even the topic here. Complete misreading. The 10 synthetic coronaviruses all have 5-8 segments marked off by restriction sites, because it's inconvenient to have much more. SC2 has 6, typical. Having 5-8 isn't weird for a natural viruses, but is has probability somewhat less than 1. (I will try to get the number.) Having any of the segments be very long turns out to be expensive and inconvenient for synthesizing viruses. So all 10 synthetic viruses have short maximal segment sizes, not quite out of the natural range but out in the tail of the natural distribution. So one picks up another likelihood factor, which I've been conservative in estimating.
I did understand the conditional probabilities. I did misread some of them, my apologies.
1) So, let's get technical. Let's suppose that P(W|L) is close to 1. We know that is not enough (recall the likelihood ratio test requires the models to be nested; these models are definitely not nested). The ratio of the likelihoods is not sufficient unless the priors are equal. There is no way that we can say that the prior of a natural origin is equal to the prior of a lab leak. Given that there has never been a leak from the Wuhan lab, I'd say the prior of a leak is very small. But, the lab has been cited for poor safety standards. So not that small. Every virus we have dealt with in human history has had a natural origin (correct me if I am wrong, but I feel safe saying the vast majority if not all have a natural origin). Thus, the prior for a natural origin is probably close to 1. Thus, if we really want to do this right, We need to take the ratio of the posteriors not the ratio of the likelihoods. Of course, the ratio of the posteriors is the ratio of the likelihoods multiplied (the bayes factor) by the ratio of the priors. So, if I grant your point that the ratio of the likelihoods favors the lab leak, I think you will have to grant me that the ratio of the priors favors the natural origins. Thus, without specific numbers, I'd say we have a wash.
I think this point of yours needs to be resigned to the trash bin
2) "The raccoon dog story fell apart. See Jesse Bloom's reanalysis." He did not really debunk the raccoon dog story, though, did he. In his own words, "But the bottom line", Bloom said, is that “when looked at carefully these data are not sufficient to conclude anything either way about whether there were infected animals.” Bloom confirmed that animals that we think are susceptible to covid infection were at the wet market. His analysis repudiated previous analysis that argued there was dog genetic material infected with covid. He argued that the genetic material we actually got from the wet market was taken far too late (a month after the initial outbreak started) to be able to conclude anything either way. Thus, I still assert your probabilities are not supported by any evidence. I think this point 2 of yours also needs to be resigned to the trash bin.
3) What are you arguing? the number for humans seems irrelevant. CGG CGG is found at low frequency in other corona viruses. Your initial claim was "P(2xCGG|L)/P(2xCGG|Z) ~100". That is we would be more likely to see the CGG CGG sequence in a lab leaked virus than a natural virus. If you are trying to argue that the 0.36 is proxy for P(2xCGG|L) and 0.03 is the value for P(2xCGG|Z), then the ratio is 12, not 100. This is strong evidence, if the priors were the same, which I have already argued they are not. However, as the paper I cited says
"Firstly, CGG CGG is part of a larger sequence—U CCU CGG CGG GC—in the gene coding for the Spike protein of SARS-CoV-2. The spike protein allows the virus to latch onto its target cells and infect them, making it crucial for SARS-CoV-2’s ability to cause disease. This 12-nucleotide sequence adds a specific molecular structure known as a furin cleavage site (FCS) to the Spike protein, which enhances SARS-CoV-2 infectivity[5,6].
While relevant for the ability of SARS-CoV-2 to infect organisms, the presence of this 12-nucleotide, FCS-coding genetic sequence isn’t suspicious. In fact, the same 12-nucleotide sequence already exists in nature: it can be found in the bat coronavirus Bat-HKU9."
Given this, we have to reassess P(2xCGG|Z). Since virus can share genetic material and there is evidence in covid-19s genetic structure that it did share genetic material with other viruses, This citation suggests that P(2xCGG|Z) should be close to 1. Given this evidence, P(2xCGG|L)/P(2xCGG|Z) < 1.
I think you need to resign this point of yours to the trash bin as well.
Final analysis, in my opinion, "P(3 synonymous SNPs|L)/P(3 synonymous SNPs|Z)= fairly big" is unjustified. But, again, you are assuming equal priors again. I don't necessarily think this point needs to be resigned to the trash bin. But, I do not think it supports your position as you think it does.
5) This was your statement: "restriction enzyme sites: P(few but no long stretches|L)/P(few but no long stretches|Z) >10" So...where is it indicated that the size is not the point??? "but no long stretches" seems to suggest size is important. But, then you said "So all 10 synthetic viruses have short maximal segment sizes" which is about the size of the segments! I am not following. You said it is not about size but then proceed to talk about the size.
Your claim this time is that synthetic corona virus have short maximal segment sizes, and they have 5-8 of them. How many do natural viruses have? How long are the segment sizes of natural viruses? You have not shared enough information for me to grant that you "pick[ed] up another likelihood factor."
Given the papers I cited, I am not going to give you this point.
Again, I reiterate my point, your numbers are either not justified or only support your position when you are performing an inappropriate statistical test.
That's very long and I have some other obligations, so I'll just start with the beginning for now.
1. I absolutely agree that the priors favor Z. I a piece I wrote 2 years ago, I used a factor of 30, acknowledging that was very rough. One of the lest few pandemics came from human error (1977 H1N1) so the priors would actually be higher than 1/30 for L/Z but H1N1 was a restart of an old virus. So it's hard to calibrate because it used t be only old viruses were present in labs but now there are many modified ones. I'm working on a draft now where I give Z 100/1 odds.
2. Jesse says that nothing can be concluded from the raccoon dog data and I agree. The absence of any identified infected intermediate host is unlikely for zoonotic origins but not extremely unlikely. I think 5 of 7 previous new coronavirus disease hosts have been identified, with the 2 unidentified being minor diseases that didn't motivate much follow-up. So this factor favors L but not very strongly.
4. I didn't actually quantitatively use the synonymous part, just commented on it.
5. Segments are the stretches between the tiny sites, not the sites! The original Bruttel paper used somewhat cherry-picked comparisons in which SC2 stood out as the most extreme of over 1000 sequences. They've toned it down to avoid assuming that the restriction enzymes relevant to SC2 would be the same ones relevant to other viruses, etc. I'm not using the full likelihood ratio that you'd get from their current analysis but just a much milder one consistent with their critics.
Bottom line- I'm on record (https://www.dailykos.com/stories/2021/6/6/2033930/-Zoonosis-or-Lab-Leak) as thinking that the priors heavily favor Z but the location heavily favors L, giving a rough wash. Then the combination of DEFUSE and sequence data, especially re FCS, give more likelihood factors favoring L. You only get to use the priors once, not multiplying them in again each time you update.
No obligations to respond here. This site ain't the best for a proper discussion. I'm sure you and I could come to some consensus on how the posteriors line up if given a chalk board, internet, and enough time.
We have moved away from my contention, though. At this point, July 2023, I'm at 50/50 as to which hypothesis is correct. I think an intentional leak has no plausibility. My contention, and the only position I'll argue for in this thread, is that, given the data available in April 2020, Andersen et alia were perfectly justified in not publishing anything but a zoonotic origin.
I was ~50-50 2 years ago, as I wrote then. I wasn't eager to change because it seemed important that people take both possibilities seriously in making policy.
I think a lot of evidence has come in since then, making it hard to keep to that position honestly. It's particularly painful to see people I like thoughtlessly repeating empty phrases, letting the likes of Rand Paul sound the like grown-ups in the room. Meanwhile dangerous GOF work is happening in Boston and Rotterdam.
As you say, I bet we could come to some approximate agreement since at least we share the basic framework, as does Nate.
>They also thought they were going to get away with it. “The truth is never going to come out ”, wrote Rambaut in one message.
This is really mischaracterizing what he's saying in that message. He wasn't celebrating getting away with a coverup, as you're saying, he was saying that IF it was a lab leak, absent concrete evidence (which was very unlikely to appear), that we'd never know that to be conclusively true.
And broadly, because a lab leak is necessarily accusatory, the evidentiary threshold is higher than the baseline, normal, zoonotical origin option. And I don't think that's really unreasonable?
I really don't have a dog in this hunt; I think both options are entirely plausible. But you really seem to be overplaying your hand on this, and the above sort of mischaracterization isn't something that I'd normally expect to see from you.
> because a lab leak is necessarily accusatory, the evidentiary threshold is higher than the baseline, normal, zoonotical origin option
Chris, I strongly disagree with this statement. In medicine and other sciences, the “baseline” is that the treatment or hypothesis one is proposing is false and evidence needs to be gathered to “reject the null.” This is done because most treatments or hypotheses tend to be later disproved.
Here there are no such baselines. It’s not as if in the past few years, thousands of deadly coronaviruses were discovered, and say 99% were natural and 1% were developed in the lab. This was a pretty novel situation.
And words like “accusatory” belong to criminal trials or politics, not scientific research.
There's no way for them to say "we think it was a lab leak" without being accusatory, though.
I get that this is contentious, and fair enough. I would have liked them to be a little more "we'll probably never be able to figure it out", personally. But I think there really is a baseline here, given the origins of MERS and SARS, and that that making an accusation -- even a weak, caveated one -- in the absence of any evidence at all, is really tough.
Mostly my beef is with the way that Nate's characterizing this though. It feels way overconfident, and bad-faith in a way I really really don't associate with Nate.
There is. They can simply say the current data is insufficient to rule out all hypothesis and we need more investigation. That's it.
What they did instead was push a narrative that lab leak is a conspiracy theory.
I don't think Andersen et Alia did that. Consider what they wrote in their conclusion:
"Although the evidence shows that SARS-CoV-2 is not a purposefully manipulated virus, it is currently impossible to prove or disprove the other theories of its origin described here. However, since we observed all notable SARS-CoV-2 features, including the optimized RBD and polybasic cleavage site, in related coronaviruses in nature, we do not believe that any type of laboratory-based scenario is plausible.
More scientific data could swing the balance of evidence to favor one hypothesis over another."
That does not sound at all like "lab leak is a conspiracy theory." In fact, that sounds a lot like "current data is insufficient to rule out all hypotheses." Though, they did say current data was sufficient to rule out an engineered virus.
You seem unaware of their starkly different behaviour in public.
https://twitter.com/abhishek_s_1/status/1682865027894898689?s=20
https://twitter.com/abhishek_s_1/status/1684637358791221248?s=20
I completely agree with the comment below from Abhishek. 🎯🎯🎯
It is unavoidably accusatory to say "we think it was a lab leak". But I think you're espousing a sort of genteel politeness ethic that says we never point the finger without a disproportionately large amount of evidence. This is dangerous, because it systematically tips us towards natural origins, and in general, if it were a norm, avoiding questions about accountability and culpability in any question or field at all.
You can ask questions about someone's culpability, implicit or not, without making direct accusations. You do run the risk of being sued, like the Data Colada folks, but somehow I think we need to find a way past that reality because it seems toxic and protects powerful bad actors.
I think it is naïve to just consider this issue in terms of good science. This was a geopolitical issue. "The science" would have had geopolitical effects. Publishing something that has no evidence to support it but will have geopolitical effects seems to demand a bit more of a cautionary approach.
Chris C has a very good point, but he goes further than he needs to. It does not matter if it is or is not accusatory. All that matters is whether China's government, the CCP, would have taken offense. I think the clear answer is "yes". When Australia called for an independent investigation into the origins of covid, the CCP blacklisted Australia. It is clear that China would have taken offense because they took offense for less. What would have China done in response? Quite possibly exactly what they did to Australia, blacklist any country that speaks up questioning the origin of covid. China accounted for upwards of 50% of PPE production at the height of the pandemic. If they had blacklisted the world, or the west, that may have actually increased the severity of the pandemic, i.e. more people, in particular medical staff, may have contracted covid and died. So, at the very least, there is a chance that publishing the uncertainty about a lab leak theory or even going on record and saying it is quite plausible could have lead to more deaths.
But, if China did blacklist a bunch of countries after some publication of the plausibility of a lab leak, what would politicians in those blacklisted countries done? Consider here in the US, we had Donnie Trump. Is it plausible that Donnie would have interpreted China clamming up when accused of leaking covid in a lab, accidentally or otherwise, that it is worse than an accidental lab leak? Maybe, a bioengineered weapon that got out and killed 1 million US citizens? The US has gone to war for far less.
So, a potential consequence of publishing the plausibility of a lab leak is basically world war 3. I think this is not the most likely outcome, but it is plausible enough that it needed to be considered.
What immediate benefits would we have realized by pushing a lab leak theory? None. Would learning that it came from an accidental lab leak helped us develop vaccines quicker? No. Would it have allowed us to find drugs that could make the disease less severe? No... well, possibly; maybe the lab would have had special knowledge. Given that they shared the genome with the world, that's likely not true. So, letting the lab leak theory out had 0 immediate benefits. The long term benefits, of course, are the integrity of science, in particular of public health, virologists, and medical researchers.
So, ultimately, I think a cost/benefit analysis based on likely geopolitical consequences of letting the lab leak theory proliferate, is squarely in favor of tabling the origins question until the pandemic was over. Scientists do not live in an ivory tower. When their uncertainty, if it got out, can only have negative consequences, table the uncertainty.
This argument is really condescending and self serving. Everyone outside the pharma industry sees the obvious need for accuracy, transparency, and honesty on these questions.
Also, if "considering the geopolitical consequences" is going to be standard for how scientific investigations are conducted, how can we ever trust science?
Not a single person in this debate is associated with the pharma industry. The authors of the paper we are discussing are all researchers at either universities or non-profit research organizations, Scripps in this case. I'm not associated with the pharma industry. Everyone in this debate sees accuracy and honesty as valuable.
Transparency is another question. About How much should I be transparent? I'm writing a paper on a specific statistical estimator. I've got hundreds of pages of math derivations, thousands of simulations, entries in my research journal recording my suspicions about what the results should look like, debates with colleagues. All of that needs to be shared? Nonsense.
What did they, in fact, hide from us? Was it evidence? No. Was it theory that could help us understand? No. They withheld their personal opinion that a lab leak was plausible; in other words, they did not share their priors with us. Was it true that, at the time, the totality of evidence supported a natural origins explanation? Yes. Was it true that there was no solid evidence that supported a lab leak explanation? Yes. The evidence they had that supported a lab leak was circumstantial, the Wuhan lab is right there, or did not distinguish between the two hypothesis, the molecular evidence was consistent with both explanations. But, they should have shared their priors with us? Why?
So, in this case, you are claiming that researchers should be transparent about their unsupported hunches. Hunches that could have significant geopolitical effects if politicians understand them incorrectly, given who the US president was at the time that was a very likely possibility. Given that everyone and their mother was citing papers and drawing drastically wrong conclusions from those papers, it also seems premature to publish unsupported hunches about a lab leak.
This issue has nothing to do with "how scientific investigations are conducted." This is about what should be published. Should we publish unsupported hunches that could cause a geopolitical storm? No.
The research into COVID was not standard. There is no reason to suppose that what we should have done with regard to a virus that was killing millions should be standard operating procedure. An analogy: consider battlefield medicine. The gold standard protocols of hospitals are likely not going to be adhered to when treating a gun shot wound on the battlefield. Does that mean we should adopt the battlefield medics procedures for our hospitals? Likely not.
Why publish what was NOT their best considered view? Acting on any one of Nate's possible motivations is wrong, damaging to science and the public' ability to gain information from scientists.
They did publish their best considered view. In particular, their position, as published in 2020, was that the lab leak hypothesis had not been ruled out but the totality of evidence made it implausible. What they did not publish was their suspicions that the lab leak explanation was more plausible than the data implied.
Scientific journals are not in the business of publishing scientists' unsupported, a prior hunches. They are in the business of publishing evidence and the conclusions that follow from that evidence. Journals leave room for scientists to speculate on further areas of research wherein unsupported hypotheses may be published. But, this is not a requirement.
Just looking at this now. It’s important to remember that Trump told the American people again and again that Covid was not dangerous, like a common cold, we would all be back in church by Easter etc. the origin issue was not important. Those were the comments that were remembered in November.
"All of that needs to be shared? Nonsense. "
Why should any of that be concealed?
There is a difference between "concealed" and "not published".
I can think of many reasons why it should not be published.
1) It will do nothing for anyone trying to understand the actual results that are supported by evidence.
2) There is so much of it that publishing it all will likely just lead to wasted paper and ink if it's printed or electricity and storage space if kept digitally. Note there are millions of scientific papers published each year. If each one of those papers required publication of all research notes, even more would go unread and thus be wasted.
3) No one will read it, especially if every paper published was also published with all the work that went into it.
4) It would take me hours and hours to collate it into something that would make sense to someone else. If every research did this, it would amount to millions of researcher-hours wasted each year.
4) There is likely to be claims amongst my research notes that are not justified by data but they are my hunches.
5) Someone might cite my unjustified hunches as the testimony of "an expert". (Notice this has happened several times over the past 5 years. Also notice, politicians and those with agendas have used the research notes and debates amongst researchers to push policies and narratives that are not justified and also to undermine science as an institution).
7) Does someone have to peer review the notes? If so, that amounts to even more researcher-hours wasted.
"I think it is naïve to just consider this issue in terms of good science"
"The long term benefits, of course, are the integrity of science"
You can argue one or the other, but that a series of geopolitical calculations for what goes on in a science journal isn't going to lead to the second thing.
I'm not exactly sure what you are saying.
I can argue one or the other of what? One of the two quotations? So, either we only consider it from good science or also consider more?
Are you saying that doing all that geopolitical calculations about what to publish is not going to lead to the long term integrity of science, the 2nd thing?
If so, well duh! That was my whole point. You have a choice: 1) maintain the integrity of science in the public's eye but possibly lead to many more deaths, and worst case a war or 2) compromise on the integrity of science but maintain relatively stable geopolitics.
Anyone that chooses (1) has twisted priorities, in my opinion.
Why couldn’t Andersen et al have published a full record of what information was available, in order to either use it to produce a probabilistic conclusion about the origin of the virus, or use it to explain why the evidence wasn’t sufficient for such a conclusion?
It seems pretty far-fetched that such a paper would cause China to ignore a gigantic PPE sales opportunity, let alone cause Trump (who was, for all his faults, relatively against sending troops anywhere) to basically end human civilization...
It seems Pretty far fetched for China to do things that it has actually done? When Japan detained a Chinese fishing trawler for violating it's maritime boarders, China shut down exports of rare earth metals to Japan, despite this being a billion dollar revenue source for Chinese miners. When Australia called for an independent investigation into the origins of covid, China proceeded along a line of first import restrictions of Australian goods and then export restrictions of Chinese goods to Australia. When a democracy activist in China received the Nobel peace prize, China reacted by shutting down imports of Norwegian fish. it is not far fetched to suppose China will do what it has been doing anytime some country offends it.
You seem to think they withheld data and evidence. They did not. They merely withheld their opinion that a lab leak, despite having no evidence to support it at the time, was still plausible.
Y'know what, while I do disagree with your wider calculus on this (so I guess I'm twisted), yes I misread you on the second sentence. So my bad, got the wording wrong.
Also I don't quite see how these particular papers had to reach the scale of the conclusion on any of this stuff in the first place. It's not like they *had* to put down it was leak or not leak. But as it played out, seems like this became an easy citation to act as if that part of the debate was finished
Not the scientists job to worry about CCP reaction. Full stop.
I categorically disagree.
It is already the case that Institutional Review Boards, IRBs, consider the potential harms and benefits of research. If the potential harms outweigh the potential benefits, IRBs will not approve a study. Thus, researchers have to consider extra-scientific considerations, in particular ethical and legal considerations.
Geopolitical risks are potential harms. Thus, IRBs will consider that. Since the researcher wants to perform their study and get it published, they have to consider potential harms that an IRB will consider. Thus, scientists will be obliged to consider the CCP's reaction when their study has geopolitical consequences.
Just because the vast majority of scientific research does not have such stark geopolitical consequences does not mean such geopolitical consequences can be ignored.
I feel there are avenues for you to object here. You could argue that the political reactions are special and IRBs are not in a good position to judge such geopolitical risks; this avenue doesn't really settle the question, though. You could argue that the geopolitical risks I outlined are very unlikely to occur; thus, IRBs can ignore them. Or, perhaps, you have another avenue of argument.
Maybe, you can actually share an argument instead of making a dogmatic statement.
I don't for one second believe that the Scientists publishing what they know not to be true was to prevent WWIII. I think the motive that Nate ascribed are more likely than what you postulate. Publishing scientific flim-flam is worse than publishing nothing at all. And for the MSM to write the findings in stone is why they are disappearing faster than polar bears.
"Scientists publishing what they know not to be true" none of the researchers that are a subject of this discussion are guilty of this.
Rather, Andersen et alia stated that the evidence cannot rule out a lab leak explanation but that the totality of evidence available in 2020 made it implausible. This was true. It was not false. They knew it to be true. Hence why they published it.
What has come to light is that Andersen et alia thought the lab leak explanation was more plausible than the data available entailed. This is not a statement of knowledge. It is a statement of personal opinion. It is a suspicion that the data is incomplete and when more complete data is made available the more complete data will show the lab leak explanation is more plausible. They chose not to publish this statement of personal opinion that was not supported by any data available in 2020. I do not see anything wrong with refusing to publish an unsupported personal opinion.
There was a huge issue, in early 2020, in getting the world to take COVID seriously. How much more seriously do you think people would have taken it if, on the first of February, global headlines had read "leading scientists believe COVID probably escaped from a Chinese laboratory", as was in fact their belief? How much more quickly and thoroughly would travel and other restrictions have been put in place? What effect would this have had on the spread of the disease? Ultimately, how many life years would have been saved? I think the answer to the last question is in the tens to hundreds of millions.
If in doubt, just tell the truth. Particularly if that's what you're paid to do. And you have no sane reason to believe that you're an expert on global geopolitics.
We've allowed the incentive structure for senior scientists to become seriously misaligned. I suppose one way of realigning it would be to follow the example of the British Navy in the 18th century - public executions of everyone involved would encourager les autres to take a different view of their self interest next time around.
I honestly don't think that would have made any difference because Covid is contagious enough (and contagious while still asymptomatic) that absent a full 100% travel blockade, it was going to make its way across every border and then spread internally regardless.
There were a couple island nations that (mostly, at significant cost) made this work; it just wasn't an option for most of the world. It was already in mainland Europe in January 2020, at that point it was going to make it just about everywhere.
It's clear it would have made some difference - it was a significant part of the success of Japan, New Zealand, Australia, Taiwan, etc. Islands, but with hundreds of millions of people. A complete quarantine of China on the first of February would have changed everything. There are valid questions about how much difference it would have made, but claiming it would have made no difference at all is silly.
Problem with that is that people who did believe it was a lab leak were the people who took it the least seriously.
The problem with tabling the discussion (and I am of the opinion both are highly plausible and we are unlikely to ever know for certain), is that tabling it essentially renders any hope of clarity completely absent, prompt evidence collection is the only hope of ever plausibly investigating the origins, and the outcome of that knowledge is far greater than simply “scientific integrity”.
It matters what the answer is because if it was inadvertently released from a Chinese lab that has huge implications moving forward for when and how we scientifically collaborate, and the world in aggregate was so deeply harmed by their casual laboratory safety measures, the world is capable of pushing on China to clean up their house. It also brings into sharp clarity the danger of all gain-of-function research and would propel increased oversight and needed regulation of these activities. Some of this may happen anyway, but the answer to the question is not idle. And given that scientists are, even now, doing things (which we know from published papers) like putting Omicron spike proteins onto Delta Strain to see if it gets worse (it does) and are carrying on this research in multiple labs around the US, and more we probably aren’t aware of, the theoretical gain from identifying the origin and recognizing risks of lab leaks from such research is all around us today and will continue to be present. This isn’t an abstract future benefit but a plausible near term benefit. We know of other instances in the last three years of COVID escaping labs where scientists have been studying it, lab leaks are more common than you might realize.
And honestly, from the Chinese perspective, they aren’t much happier about their wet markets being blamed for the origin than their lab. It might seem to you as if they’d strongly prefer one, but they don’t actually like either one of them, and have obfuscated evidence for both.
No one is calling for tabling the question indefinitely. Rather, table the question until the pandemic is handled. The question is valuable, in the long term, for exactly the reasons you mentioned in your second paragraph. But notice, all those reasons are long term. None of them would have helped save lives in 2020.
You are right. Tabling the question until later may cause data collection problems. But, I'm less worried about that. We have genetic structure of the early COVID variants. China, presumably, has as good a record as possible of the the spread in Wuhan. There is not much more you can use that plausibly would have been available back in 2020.
Your last paragraph has merit but is also wrong. Sure, they don't like either. But, they like the lab leak explanation far less than the natural origins in the wet market explanation. They blacklisted Australia because Australia basically suggested that a lab leak explanation was plausible. They did not blacklist anyone for pushing a natural origins at the wet market explanation.
There is a sort of baseline on recent pandemics from straight zoonosis (several) and from lab leaks (1977 H1N1). So there are priors favoring Z. But the aren't infinite. Even at 100/1 priors it looks like location, DEFUSE, the sequence data, probable discovery of ancestor in Vero cell cultures, near absence of lineage A from HSM, still no trace of intermediate wildlife host,... will swing the posterior odds to favor L.
It's probably best if governments keep mumbling about inconclusive results, but meanwhile we need international regulation of GOF, especially in places like Boston and the Netherlands, where some really crazy shit has been done.
Really, really badly mischaracterizing it, to an indefensible degree. No excuse for it. Nate should correct.
> And broadly, because a lab leak is necessarily accusatory, the evidentiary threshold is higher than the baseline, normal, zoonotical origin option. And I don't think that's really unreasonable?
That could be a reasonable answer if Proximal Origins had withheld judgment and said that the evidence is inconclusive. But that's not what the scientists said: they told us we *could* draw a conclusion, and that the evidence likely ruled out a lab leak.
And the paper did lead to accusations, just not against China, but rather against the American "racists" and "conspiracy theorists" who defended the plausibility of a lab leak. If the organizing principle was to avoid accusation, why didn't scientists stand up and defend people against these attacks, given that they privately believed a lab leak was plausible?
Oof, this has really dropped my trust in Nate. Thanks for pointing it out.
By the way, for anyone who wants to check the full context for themselves, you can see it in the tweet Nate linked (https://twitter.com/mbalter/status/1679098392587255809/photo/1), or go to page 8 of the first PDF in the leak (https://public.substack.com/p/covid-origins-scientist-denounces).
I think it reads like someone who privately understands what the others are saying about strange “coincidences” and such, but also understands just how bad it would be for his field if the lab leak hypothesis were to gain traction, so is trying to shape the conversation toward dismissal by pointing out there will never be conclusive proof of a lab leak -- he’s saying it’s a safer bet to exaggerate their confidence in a zoonotic origin.
Thank you for the links!
I would characterize myself as strongly favoring the lab leak hypothesis and strongly believing there was a cover-up. I clicked through and read the source, and came to the same conclusion as Chris -- Nate badly mischaracterized this quote. The scientists are not saying what he accuses them of. He should correct this.
I'm sad to say I have seen Nate take highly implausible (to me) readings of sources before. At this point, I don't especially regard "reading text and generating unbiased analysis of what the people generating it meant" high among his skills. (I think that's a more likely explanation than that he is being deliberately unfair or biased).
100%. I read that line, blanched at it, and jumped to the comments to see if I was overreacting.
It's really clear in context that he means "*we* will never know the truth", not that they know the truth and are covering it up.
Sheesh, this is really bad.
Reading the source material, I agree with Chris here. Nate mischaracterizes this quote.
I agree with Nate’s greater point, but this mischaracterization undermines the credibility of it.
Where did Nate state or imply that the authors were "celebrating" something? Their mindset was that the deception would stand, and "The truth is never going to come out" was part of that mindset.
They're acknowledging what they think the truth is and that there will be no conclusive evidence, therefore opting for the safer, more opportunistic, career-oriented approach. Tbh I'm not sure why they're getting defended so hard here.
If you're going to accuse someone of mischaracterization, don't mischaracterize what they have said. Nate didn't say that Rambaut was celebrating getting away with the cover-up. He said that Rambaut and others thought they would get away with the cover-up. He backed this up with a quote from Rambaut which seems to me to provide evidence for that assertion.
You, on the other hand, don't have a quote providing evidence for your assertion that Nate was saying that Rambaut was celebrating the cover-up. That's because Nate didn't say it.
I would normally think that your misunderstanding might be linked to you already having strong preconceived views on this subject, which have lead to you searching for a way to discount the damning evidence Nate sets out. But reading on, I see that you say that you don't in fact hold such views, so I must be mistaken.
I directly quoted Nate, and obviously the whole post is up there for people to read. I think it's very clear that Nate's attributing that specific quote, "the truth is never going to come out", to them "thinking they could get away with it", meaning Nate's asserted cover-up.
But read the actual context of the quote; that's not at all what Rambaut is talking about.
Rambaut is talking about the truth not coming out. Which, on the plain reading of what Nate wrote, is what Nate is saying Rambaut was talking about. You read something into what Nate wrote that wasn't there.
Lets give a full quote of what Nate says since that is in dispute:
> They also thought they were going to get away with it. “The truth is never going to come out ”, wrote Rambaut in one message.
That is very clearly accusing Rambaut of being dishonest here, of thinking that lab-leak was plausible and much more likely than Rambaut had published. The rest of the article is accusing them of that too.
But had Nate included the very next words from Rambaut in that message "(if escape is the truth)" he would've, and we would've immediately seen that the context absolves Rambaut of that accusation at least in that message. It's a discussion of a hypothetical of how *in the world in which COVID started as a lab leak* there wouldn't be evidence available to prove it (due to particulars of that situation, of China's secrecy, whatever).
There really isn't a valid argument to the contrary, Nate was misleading on presenting that quote from Rambaut and should be correcting his article.
No. He accuses Rambaut and his co-authors of being less than honest - but not here. He accuses them of being less than honest with reference to the private messages that show that their public statements misrepresented their actual beliefs.
Here, he states that Rambaut thought that the truth would not come out. Which is only suggestive in combination with the other evidence. By itself, it means nothing.
Anyone would think that those fixating on a simple, accurate, and appropriately sourced statement of relatively minor importance were doing so because they desired an excuse not to have to think about what is staring them in the face.
Nate didn't say they were "celebrating" a cover up. He said they thought they would get away with publishing a purposefully misleading article. Clearly, they were confident in the deception, although KA worried about what the spooks might know.
Also, accepting lab leak would result in stronger safety policies. For that, the evidentiary threshold should not be so high.
No. He’s saying lab leak is plausible. But he feels he can lie about that because he will never be proven wrong definitively.
And taking your two assertions together (higher evidence threshold & high unlikelihood of concrete evidence) it would be nearly impossible to suggest a lab leak leak had occurred under any circumstance.
If you read the linked articles, the 3 scientists working on weaponizing covid viruses are named and were the first to die.
You mean these people? https://www.science.org/content/article/ridiculous-says-chinese-scientist-accused-being-pandemic-s-patient-zero
Yes, those people
https://public.substack.com/p/biden-is-breaking-law-by-not-releasing
It was absolutely a deception, because it allowed them to start the conversation with zootropic as the null.
I really do think that there was a hard turn after Trump was elected by center-left media outlets to fight fire with fire, meaning if right wing media outlets were going to lie in order to push an agenda, then they are going to do it to. That's not to say it never happened, but Trump getting elected made it even more prominent. In this instance, because 'Trust The Science' was the motto being pushed, anything that might undermine that message wasn't reported on. Needless to say I think this is an extremely bad direction for mainstream media outlets to go.
The old line (think Bush era) used to be that the left lacked a Fox News, and was kind of boxed in by journalistic norms whereas Hannity could do what he does every night. There was a point there on several levels - however, the internet changing the nature of media platforms and a reliable college-educated audience has made that something of the past. And yeah, creates some *very* bad incentives.
Except that line was false. Every mainstream outlet was the left's Fox News.
It was, and is, all about Trump. Anything deemed helpful to him, in any way, must be discredited, sneeringly dismissed and suppressed. Anything that is thought to hurt him must be propped up, repeated, treated as gospel truth - no matter how shaky it really is.
It feels like the left, having been fairly defensibly the "party of science" for the past few decades, has learned the lesson that they don't actually have to have beliefs that align with science or even do science at all - they can just skip to the end, proclaim loudly that "the Science has spoken", and that any other position is a conspiracy theory not worthy of a moment's attention.
They've been doing that for decades, congratulations on finally noticing.
With Biden admin (finally) cutting funding to WIV, it seems the slow walk from "lab leak is a racist conspiracy theory" to "well, shoot, we don't know what happened" to "it was a lab leak but it would be impolite to say so" is well underway.
Biden himself always seemed like a quiet lab leak believer. I think one thing that really pushed scientists against the lab leak theory is that gain of function research was very contentious and it wasn't until 2017 when it was allowed in the US. So to have a possibility that gain of function research had the worst result possible only 3 years into it would lead a lot of scientists to really want to hammer home the zoo tonic origin and ignore anything that might point in the other direction. Hell, it seems pretty clear that covid wasn't a modified virus, but even the optics of where it escaped from would put gain of function research in the crossfire
One thing that some in the media did was try to conflate the "it's bio terrorism from China" theory with the lab leak theory. It was pretty unlikely and remains unlikely that this was an intentionally bioengineered pandemic, but early on the lab leak theory was treated the same way.
Why, in your opinion, does it seem pretty clear that COVID wasn't a modified virus?
Biden has been lab-leak curious from the beginning actually
Kristian Andersen has stated many times publicly that (1) he initially believed that a lab leak was likely, (2) he subsequently changed that view based on the scientific considerations that are detailed in his scientific publications, and (3) he has always believed, and has always written, including in the Proximal Origins paper, that a lab-leak scenario is possible (though he now doesn't believe it's plausible). He explained this in detail in an all-day interview, half of which was conducted by a House Republican staffer, as well as in his prepared testimony:
https://oversight.house.gov/wp-content/uploads/2023/07/2023.06.16-Andersen-Transcript.pdf
https://oversight.house.gov/wp-content/uploads/2023/07/Testimony-of-Dr.-Kristian-Andersen.pdf
Nothing in Nate's post or in the links contradicts Dr. Andersen's version of events. In fact, none of them of even attempt to. Nate quotes Dr. Andersen's February 1, 2020 email as if it were a smoking gun that Dr. Andersen was lying, without even addressing the fact that Dr. Andersen has repeatedly stated that he changed his mind between February 1, 2020 and the time of publication.
I mean it's obvious that Dr. Andersen did change his mind at some point! He has been publishing papers to the effect that the lab leak is unlikely continuously since early 2020. What is the argument here? That his initial impressions from February 1, 2020 were his true and final feelings and that everything he has said and published since then is disingenuous? That's absurd.
To be clear, Dr. Andersen could well be wrong. I take no position on how plausible the lab-leak hypothesis is, because I don't have the training necessary to evaluate the scientific evidence. Plenty of people who do have the training take different views of the evidence and that's fine.
But I do have the training to evaluate allegations of fraud. And I see no persuasive evidence of it here. Until I see a careful, good-faith assessment of the record that considers Dr. Andersen's detailed explanations and attempts to refute them, I consider the allegations of scandal completely unproven. And this post, which asserts that the facts are "simple and clear" without even engaging with Dr. Andersen's public statements on the issue, and mischaracterizes the statements that it does cite (see Chris C's comment) gives me zero confidence that Nate has made a good-faith effort to get to the truth.
> he has always believed, and has always written, including in the Proximal Origins paper, that a lab-leak scenario is possible (though he now doesn't believe it's plausible)
In April the Slack shows Andersen privately worried that he can't disprove a lab leak. He even posits a potential mechanism by which it could be engineered:
> [Andersen]: We also can't fully rule out engineering (for basic research) - yes, no obvious signs of engineering anywhere, but that furin site could still have been inserted via gibson assembly (and clearly creating the reverse genetic system isn't hard - the Germans managed to do exactly that for SARS-CoV-2 in less than a month).
How does this statement square with Proximal Origins, which says "we do not believe that any type of laboratory-based scenario is plausible"? He just posited a scenario by which it could happen! This scenario does not appear to rely on any kind of fantastical black swan event -- quite the opposite, it says that the Germans did exactly that in less than a month. I don't know how to read this statement as anything other than "an engineered virus is plausible."
If this is going to come down to weasel words, where "implausible" can mean "I privately admit that it could well have happened that way," then they are entitled to their weasel words but they have lost my trust.
1. Implausible and "can't fully rule out" are consistent with each other, as you acknowledge. In fact, the March 2020 paper expressly said that lab leak hadn't disproved. You can call it weasel words, or you can call is "accurate use of the English language."
2. When all of someone's private communication on a subject have been subpoenaed and publicly released, someone attacking that person based on a single message has the burden of proof of showing that the document is being read fairly in context. There has been no effort to do that here; there is no mention at all of the context of the message or what came before or after. Before I can even begin to criticize Andersen, I'd like to see someone analyze the full record and show how it contradicts Andersen's public statements. That hasn't been done.
3. Nate and others are accusing Andersen of fraud. That is an extraordinary claim and nobody has come close to proving it or even seems particularly interested in doing so in a careful and fair-minded way. If, after all of someone's private communications have been publicly released, the best that you can do is identify one isolated message and assert that is shows the scientist used "weasel words," this strongly suggests the fraud accusations are baseless.
I agree that fraud is a strong accusation, and based on what I’ve seen so far I don’t know if I would go that far.
That said, the authors’ main defense is to hide in the ambiguity of what the word implausible means. When their colleagues are citing their paper as evidence that Covid is zoonotic, and signing letters condemning “conspiracy theories” that it could have been a lab leak, then “implausible” is taken to mean “we can dismiss doubters.” But when their own private doubt is made public, then “implausible” is taken to mean “we were always open about the fact that we hadn’t ruled it out completely.”
They can’t have it both ways. If the most expert scientist is allowed to have private doubts, then everyday people are entitled to have public doubts.
If scientists are resorting to this kind of argument then their entire enterprise is toast. The public will not accept arguments of the form "scientists speak a very special form of English in which nothing they say can be taken at face value regardless of how clear it may seem". Why fund people who make such arguments?
There's a lot of context about the messages available. Nate Silver is letting you know more messages exist, it's on you to go read Racket and Public if you want to see more of their logs. But be aware: it doesn't get less damning, it gets moreso.
They just lied. Yes, they engaged in a kind of fraud. If they were doing this sort of thing in the private sector and about accounting instead of deadly viruses they'd be fucked already and probably on the way to jail. Why is it so hard to accept that? It's called lying, not "accurate use of English". Academics are not morally purer than normal people, frankly the evidence is strong that they're dramatically more corrupt than normal people are. It's a culture that doesn't seem to have any norms of honest behavior. Perhaps unsurprising given its essentially communist setup.
Why demand that scientists, among themselves, contrive to speak so every possible raging conspiracist dumbass is pre-emptively neutralized? And no, you don't have the full context. You just want to believe you do.
Yes, the populist attack on science and education has blinded liberals to the truth about scientists. They are human beings like everyone else. They have their personal stakes, their preconceptions, their ambitions, and their blindnesses. This is not a new thing, it has always been thus. (If you doubt this, you ought to read Charles Mann's 1491 and follow the endless roadblocks scientists threw up along to the implications of archeological finds that shed light on the history human beings in the Americas.)
Scientific inquiry is an incredibly powerful way to seek the truth of many matters. It has shed light on myriad things, and will continue to do so. But it is always a work in progress, and can never be personified as the final word. To do so is to treat science as your religion -- which is exactly what many people do. Further scientific inquiry will upend a fair portion of what appears true now, and it will update and adjust a great deal more. And that would be true, even if individual scientists didn't have all the personal weaknesses shared by the rest of our species.
As a physician who has followed your work across domains for about 15 years I’m astounded by the how wrong have gotten this Nate. Sad to see your name behind this weak, conspiratorial to be point of being delusional “reporting”. 538 is worse off without you but you seem to be missing your edge now too. Unsubscribing but maybe will circle back in a year or two.
It would be great for you to break down why Nate is wrong instead of simply asserting it. The lack of public back and forth on this is extremely frustrating.
Thank you for asking this Petey. My question exactly. If someone thinks Nate is obviously wrong, please explain why.
Long read, but I explained many of the reasons in a blog post:
https://medium.com/microbial-instincts/the-case-against-the-lab-leak-theory-f640ae1c3704
That's a little bit out of date, I've found more arguments in favor of natural origin since, but it should give you an overview, and you can dig into the linked sources if you need more.
I also agree that it would be nice to have more public debate on this. I'm actually in negotiations with Rootclaim to debate this for cash, though it's still TBD whether we'll agree on favorable terms and when the debate will happen.
Nate is wrong because he's daring to question the holy prophet of science, Fauci.
And I'd guess you are wrong because you've gotten all your information about Dr. Fauci from grifters and kooks.
Where by "grifters and kooks" you mean anyone who dares to question the Holy Prophet Fauci.
I think I mean you, too.
As I was saying.
Wow, you seriously think, with all the data about ) gain of function research being done in Wuhan, 2) the fact that years later, no animal source for this leak has been found (as compared to other, truly natural outbreaks), 3) the ONI report, and 4) even comments from the scam letter signatories that it looked engineered, that it's a natural outbreak? Astounding.
1) Do you have any idea how many places do GOF? Without that, it's impossible to evaluate the importance of this point.
2) Animal sources are almost never found viruses. We've been looking for an origin for MERS for a decade. This "Middle East" RS appears (possibly) to have come from bats in the vicinity of South Africa, thousands of miles away, made a leap to camels, and then to humans.
3) The report says nothing about origins, and doesn't present strong evidence for a lab origin. For example: "We continue to have no indication that the WIV’s pre-pandemic research holdings included SARSCoV-2 or a close progenitor, nor any direct evidence that a specific research-related incident occurred involving WIV personnel before the pandemic that could have caused the COVID pandemic.'
1) So what? that doesn't mean that a) it's a good thing of b) it wasn't happening in Wuhan (it was), and didn't escape.
3) Otherwise known as CYA....
So what? If I tell you that every city in China has a GOF lab, it wouldn't change your view?
Don't reply if you don't want to use logic and common sense. You're using the phrase the phrase GOF like it's a boogeyman.
CYA? Then why did you even mention it?
If every city had a GOF lab with DEFUSE-like plans for coronaviruses, that would certainly change one of my Bayes update factors, just as you say. It's highly unlikely that there are more than several, if any, others. There are no reports of repeated international warnings abut sloppy labs other than in Wuhan. Meanwhile, there are >20k wet markets, with only about 4 in Wuhan, so my simple population-based x100 based on Wuhan location is probably an underestimate.
I would love a citation for ANY of what you just wrote.
Doesn't matter. Still doesn't mean that that wasn't how the virus was created and didn't escape.
I was referring to your cherry-picked quote from the report.
Sorry, "reason and logic" and "people who still believe COVID19 was natural" is an oxymoron.
"Cherry-picked quote" literally from the report you tried to cite as evidence which DOES NOT CLAIM WHAT YOU SAID IT CLAIMED.
If you had an argument, you wouldn't need to lie.
1 is completely irrelevant. If the outbreak started in Wuhan then all that matters is what mechanisms were available in that location that could have caused it.
2 is also completely irrelevant if substantially more time/money has been spent in this case vs other cases. MERS was not a pandemic so it would hardly be surprising that less has been spent looking for it’s origins. Of course you’re free to provide evidence to back up your claim.
Re your point 2- the proximal hosts (camels and either civets or raccoon dogs) of MERS and SARS CoV1 were found. The proximal host of SARS CoV2 has probably also been found in sequences from Vero cell cultures sequenced early at the Sangon lab, with 3 SNPs differing from original SC2, all in the direction of the ancestral sequences.
It's difficult to overstate the damage done over the past three years to "trust the science" and "trust the experts" by these people. Before COVID I was firmly in the camp that thought the scientific consensus was usually correct, or at the very least had some merit. The climate change debate had weakened that trust somewhat already, but a lot of the problems there I had blamed on an overzealous media addicted to sensationalism and pushing an agenda. Now? That willingness to give experts the benefit of the doubt is gone.
Yes the real mind-bender is when you start to update past beliefs in light of this "new" (not really new) information. Climate stuff is not different to COVID it seems. The ClimateGate emails were that field's lab leak coverup equivalent, but because our society worships academia it was all swept under the carpet and nothing was done even when they were writing the most damning things.
I am utterly unshocked that you think the 'ClimateGate' emails were evidence of truth about climate being covered up. It wasn't swept under the carpet, it was investigated at tedious length. You're just mad because no evidence of such a coverup was found. You probably think 'trick' meant 'trick' the public', too.
One should not forget that Trump's correct statement that the Wuhan virus came from Wuhan was used as a stick to beat him by the media:
ABC:
“Trump's 'Chinese Virus' tweet helped lead to rise in racist anti-Asian violence”
Mar 18, 2021 — "Anti-Asian sentiment depicted in the tweets containing the term 'Chinese Virus' likely perpetuated racist attitudes and parallels the anti- ...
UCSF:
Trump's 'Chinese Virus' Tweet Linked to Rise of Anti-Asian Violence
UC San Francisco
https://www.ucsf.edu › news › 2021/03 › trumps-chinese...
Mar 18, 2021 — In the week after former President Donald J. Trump tweeted about “the Chinese virus,” the number of coronavirus-related tweets with ...
Trump's 'Chinese Virus' Tweet Helped Fuel Anti-Asian Hate ...
Forbes
https://www.forbes.com › roberthart › 2021/03/19 › tr...
Mar 19, 2021 — Amid an uptick in anti-Asian violence, Trump's rhetoric has been singled out as particularly damaging.
How Trump Fueled Anti-Asian Violence in America
The Diplomat
https://thediplomat.com › 2021/06 › how-trump-fuele...
Jun 8, 2021 — Trump's encouragement of white supremacy and his strident anti-China rhetoric proved a toxic combination for Asian Americans.
Racist anti-Asian hashtags spiked after Trump first tweeted ...
Washington Post
https://www.washingtonpost.com › nation › 2021/03/19
Mar 19, 2021 — A group of researchers found Twitter users published more anti-Asian hashtags the week after he first tweeted about the “Chinese virus.”
Uh, how was it not? Because "Chinese virus" and "virus that came from China" sound the same to me.
Yeah, Trump was merely being informative and accurate. /eyeroll/
As a scientist myself I'm kind of of two minds here. I think one major issue is that most of us are used to writing for other scientists to read, and so because we're all humans, I think a lot of scientists are trained to sort of correct for those sorts of biases that we know exist. That's not to say that the authors acted ethically--they obviously didn't--but I read a lot of papers that I think are crap but we just sort of say "I think that's crap" and then move on, or if we're really motivated, write a paper refuting it. We're trained very well for that, so writing a crappy paper isn't usually seen as a big scandal, it's just seen as sloppy work. What we're not trained for is talking to the general public. And for most of us, that's fine. In my situation, the chance of there ever being a global emergency relating to non-classical nanocrystal growth is rather low, so it's unlikely that there will aver be any sort of public policy or what have you based on my work. But it seems like for people in public-facing fields, that should be emphasized. I realize it can be difficult to determine sometimes what constitutes a public-facing field, but I feel like something like "public health" where "public" is in the name might fall into that category. So from that standpoint, even with as charitable as possible an explanation for their shoddy work, as public health professionals they need to understand that the general public is not trained to critically analyze science papers but is constantly primed to think about politics, so they need to be extra careful about controlling for political bias in their work if they know that the work is going to be of public interest.
I think there's also something to be said from the other direction as well--the media is absolute crap at reporting science. Before COVID that was sort of just something that was irritating but since COVID it's really exposed how much of a problem that is. The reality is that interpreting science is hard and those of us whose job it is to do that go to school for a very long time to be able to have the technical skills for that. Ultimately due to media biases they'll often just report on the things they already wanted to report on and pick scientific studies to cite that support what they already believe. The secret is that scientists do that too. We're all human and even though we try to control for our own biases, we do get excited about our own work and want to find references that agree with us. The hope is that if something bad was missed, that will get sorted out in peer review, and then if it doesn't the people who read the paper will read it critically. One thing I wonder about is that in the early days of COVID, a lot of journals instituted policies that would fast-track COVID-related papers for review, which was a policy that obviously made a ton of sense at the time but also likely led to overworked reviewers who likely had COVID-related research of their own and their lives, like everyone else's, were uprooted and all that, so I wonder if a lot of these reviews got kicked to postdocs and grad students who haven't reviewed a lot of papers before and let more slip through the cracks, and then what slipped through the cracks made it to the media, and then Pandora's box was open. As you pointed out, early COVID was a hard time for everyone, but I think that for those people whose job it was to provide and disseminate the best and most current information, including the public health scientists, the journal reviewers, and the media, there were a lot of failures that compounded and led to serious problems such as this one.
“the media is absolute crap at reporting science.”
It’s pretty safe to posit at this point that the media, with a few exceptions, is absolute crap at reporting almost anything.
Education polarization has made them overly trusting of certain sorts of authority figures, most especially academics/university administrators, non-profit workers, and civil service workers, and overly skeptical of others, most prominently police and businesspeople.
But that’s a small part of the problem which encompasses only a small fraction of stories and reporting.
The major driver is that “reporter” is no longer a job with the combination of pay and prestige, in most instances, to attract and retain top-notch talent. To be frank, most reporters I know or have read enough of to have some familiarity with are complete mid-wits with neither the raw intelligence, nor the work ethic, to actually analyze and critique almost *any* complex narrative in almost any field of endeavor.
And because most don’t actually have the ability to do this, the vast majority of reporting has become essentially editorial in nature, with a reporter’s meager “analysis” driven almost solely of their priors towards the actors involved.
Thanks for chiming in. It's good to hear from scientists and researchers when things like this happen.
Perhaps Nate went too far in characterizing the scientists thoughts on the Nature article. We don't know what people think. But Nature made sure to publish an article that loudly said COVID wasn't a lab leak. Frankly, scientists are loosing ground on trust from the public. Covidgate, p-hacking, faux scientific papers being published, before that was Climategate, etc. -- these instances feature peer reviewed literature and journals that are at least suspect if not part of the problem. We can waft some of these aside in some cases but in COVID's case it's Nature and Lancet, the premier publishers, so the problems with scientific literature gaining trust from the public has hit rock bottom.
Trust can't be gained from simply training scientists to speak more eloquently to the public. The scientific publishing process needs to divorce itself from whatever has tainted its well.
I support open science. But that would get rid of all the money that flows via the journals. Perhaps that money is the problem.
Wife and I were talking about that this morning, in this context and as it pertains to medical research with respect to nutrition.
I have no time or patience for physicians and scientists who bemoan the public’s lack of trust in science. Do they think we’re blind to the influence of money on what they call “science”? They work for institutions that whore them out to political and commercial interests in exchange for money and political credit, publish their work in journals that whore themselves out for money and political credit, grant their blessing to others’ shoddy work and attack work that challenges the mainstream narrative, again for money and political credit.
Having whored themselves out, they stand in front of the world and try to present themselves as “impartial scientists”? And bemoan our “lack of faith” or “lack of trust”? Faith and trust have no place in the scientific method, nor has the scientific community conducted themselves in a manner worthy of trust and faith in their collective integrity.
At the very least, Andersen, et.al. went into their investigation with a definite preference for the preferred outcome. They expended a great deal of effort looking for evidence that would support a plausible rationale for natural evolution, but did they work as hard to establish a plausible rationale for laboratory evolution? The content of their Slack channel suggests not. They wanted to prove one hypothesis, and disprove the other. I don’t believe they accomplished either goal.
The truthful conclusion of their investigation, such as it was, would have been “we simply can’t reach a definitive conclusion”. But that wasn’t the outcome they wanted, so they wrote the conclusion they and their patrons wanted and massaged their paper to support it. It wasn’t honest science, but it was very much in line with “science” as we see it conducted in the public sphere today.
Can I suggest that folks read the documents for themselves. Very clearly shows that they revised their paper to be more skeptical of lab leak after new data came to light in March. Prior to that, some were personally leaning towards lab leak without hard evidence for it, and being careful to distinguish publicly between what they reckon and what they can scientifically conclude. Very surprised to see Nate Silver swallowing the misleading framing.
Here's brief partial list of observed features that contribute substantial likelihood ratios to update the prior odds that favored Zoo. All the updates favor LL.
1. location: P(Wuhan|L, FCS) ~1, P(Wuhan|Z, FCS) <0.01 ratio >100
2. Intermediate hosts found: P(Vero cells but no wildlife|L)/P(Vero cells but no wildlife|Z) = big
3. FCS sequence: P(2xCGG|L)/P(2xCGG|Z) ~100
4. FCS neighborhood: P(3 synonymous SNPs|L)/P(3 synonymous SNPs|Z)= fairly big
5. restriction enzyme sites: P(few but no long stretches|L)/P(few but no long stretches|Z) >10
etc.
So after talking to experts I would now downgrade
Intermediate hosts found: P(Vero cells but no wildlife|L)/P(Vero cells but no wildlife|Z)
from "big" to ">1 but not huge" due to the messiness of the Vero sample data.
I'm going to take issue with all of your numbers. I presume Z=zoonotic origin; L=lab leak; FCS= furin cleavage site (?).
1) P(Wuhan|Z, FCS) is definitely not less that 0.001. It is also not the case that P(Wuhan|L, FCS)~1. The first stems from the fact that we know wet markets do make for solid breeding ground for virus that can jump species. The second stems from the fact that researchers and staff at the Wuhan lab interacted with people from all over China and the world. An asymptomatic research that infects his cousin who then travels back to Beijing is very plausible. Both of your numbers are questionable.
2) This point is irrelevant considering that COVID has been found amongst the genetic material of dogs recovered at the wet market. Thus, P(vero cells but no wildlife)=0.
3) The 2xCGG sequence is found in nature amongst other viruses. Thus, your estimated probabilities are suspect. https://healthfeedback.org/claimreview/the-cgg-cgg-genetic-sequence-and-furin-cleavage-sites-also-exist-in-naturally-occurring-viruses-these-features-arent-evidence-of-genetic-manipulation/
4) This paper suggests that synonymous SNPs in viruses is a way for the virus to escape the hosts immune system. If true, that would entail there would be an adaptive advantage to having synonymous SNPs. Thus, synonymous SNPs would not be unexpected. https://pubmed.ncbi.nlm.nih.gov/31416386/
5) This one doesn't make any sense to me. Restriction enzyme sites are small sequences of code that bacterial enzymes bind to and cleave. Everything I've found say these are generally between 6-8 base pairs long. Thus, the probability that there would be few of them and they would be small would seem to be quite large for any virus, irrespective of origin. Thus, these numbers are suspect too.
In summary, all the numbers you shared are highly suspect.
Wow, a combination of misreading and misunderstanding conditional probabilities.
1. I used P(Wuhan|Z) of 0.01, not 0.001, although based on the distribution of wet markets one might justify the lower number. Your visitor story is just a reminder that P(Wuhan|Z) is not zero. But conditional on Z Wuhan is no more likely than other places with similar population, maybe less. Conditional on L, sure P(Wuhan|L)<1 because there are other labs working on coronaviruses, but this is a big one, known to be sloppy, known to plan FCS insertion. So P(Wuhan|L) is not a whole lat less than 1.
2. The raccoon dog story fell apart. See Jesse Bloom's reanalysis. SC2 RNA negatively correlated with stalls with potential hosts, other than humans. But it's true that on closer examination the samples with Vero cells and SC2 RNA are not a smoking gun, because that RNA could have come from humans rather than the culture cells.
3. The CGG numbers for this group of viruses are solid, ~0.03. For humans, ~0.36. There's nothing wrong with this calculation except that I was conservative in treating the 2 sites as independent under the L hypothesis.
4. Synonymous mutations are indeed common. They just seem way too common here, e.g. one stretch of 60 mutations, all synonymous.
5. Wow, the size of the restrictions sites is not even the topic here. Complete misreading. The 10 synthetic coronaviruses all have 5-8 segments marked off by restriction sites, because it's inconvenient to have much more. SC2 has 6, typical. Having 5-8 isn't weird for a natural viruses, but is has probability somewhat less than 1. (I will try to get the number.) Having any of the segments be very long turns out to be expensive and inconvenient for synthesizing viruses. So all 10 synthetic viruses have short maximal segment sizes, not quite out of the natural range but out in the tail of the natural distribution. So one picks up another likelihood factor, which I've been conservative in estimating.
I did understand the conditional probabilities. I did misread some of them, my apologies.
1) So, let's get technical. Let's suppose that P(W|L) is close to 1. We know that is not enough (recall the likelihood ratio test requires the models to be nested; these models are definitely not nested). The ratio of the likelihoods is not sufficient unless the priors are equal. There is no way that we can say that the prior of a natural origin is equal to the prior of a lab leak. Given that there has never been a leak from the Wuhan lab, I'd say the prior of a leak is very small. But, the lab has been cited for poor safety standards. So not that small. Every virus we have dealt with in human history has had a natural origin (correct me if I am wrong, but I feel safe saying the vast majority if not all have a natural origin). Thus, the prior for a natural origin is probably close to 1. Thus, if we really want to do this right, We need to take the ratio of the posteriors not the ratio of the likelihoods. Of course, the ratio of the posteriors is the ratio of the likelihoods multiplied (the bayes factor) by the ratio of the priors. So, if I grant your point that the ratio of the likelihoods favors the lab leak, I think you will have to grant me that the ratio of the priors favors the natural origins. Thus, without specific numbers, I'd say we have a wash.
I think this point of yours needs to be resigned to the trash bin
2) "The raccoon dog story fell apart. See Jesse Bloom's reanalysis." He did not really debunk the raccoon dog story, though, did he. In his own words, "But the bottom line", Bloom said, is that “when looked at carefully these data are not sufficient to conclude anything either way about whether there were infected animals.” Bloom confirmed that animals that we think are susceptible to covid infection were at the wet market. His analysis repudiated previous analysis that argued there was dog genetic material infected with covid. He argued that the genetic material we actually got from the wet market was taken far too late (a month after the initial outbreak started) to be able to conclude anything either way. Thus, I still assert your probabilities are not supported by any evidence. I think this point 2 of yours also needs to be resigned to the trash bin.
3) What are you arguing? the number for humans seems irrelevant. CGG CGG is found at low frequency in other corona viruses. Your initial claim was "P(2xCGG|L)/P(2xCGG|Z) ~100". That is we would be more likely to see the CGG CGG sequence in a lab leaked virus than a natural virus. If you are trying to argue that the 0.36 is proxy for P(2xCGG|L) and 0.03 is the value for P(2xCGG|Z), then the ratio is 12, not 100. This is strong evidence, if the priors were the same, which I have already argued they are not. However, as the paper I cited says
"Firstly, CGG CGG is part of a larger sequence—U CCU CGG CGG GC—in the gene coding for the Spike protein of SARS-CoV-2. The spike protein allows the virus to latch onto its target cells and infect them, making it crucial for SARS-CoV-2’s ability to cause disease. This 12-nucleotide sequence adds a specific molecular structure known as a furin cleavage site (FCS) to the Spike protein, which enhances SARS-CoV-2 infectivity[5,6].
While relevant for the ability of SARS-CoV-2 to infect organisms, the presence of this 12-nucleotide, FCS-coding genetic sequence isn’t suspicious. In fact, the same 12-nucleotide sequence already exists in nature: it can be found in the bat coronavirus Bat-HKU9."
Given this, we have to reassess P(2xCGG|Z). Since virus can share genetic material and there is evidence in covid-19s genetic structure that it did share genetic material with other viruses, This citation suggests that P(2xCGG|Z) should be close to 1. Given this evidence, P(2xCGG|L)/P(2xCGG|Z) < 1.
I think you need to resign this point of yours to the trash bin as well.
4) Again, if there was an adaptive pressure to have synonymous mutations, we might get large strings of synonymous mutations. But, this point of yours is not very convincing. This paper suggests that covid has fewer synonymous mutations than other RNA viruses. It also has several sites where synonymous mutations predominant. https://virologyj.biomedcentral.com/articles/10.1186/s12985-023-01982-8#:~:text=Mutational%20analysis%20showed%20that%20SARS,synonymous%20mutations%20were%20overwhelmingly%20predominant. This article suggests, as my previous one did, that the synonymous mutations are an adaptation to the human immune system. https://academic.oup.com/ve/article/8/1/veac026/6553895.
Final analysis, in my opinion, "P(3 synonymous SNPs|L)/P(3 synonymous SNPs|Z)= fairly big" is unjustified. But, again, you are assuming equal priors again. I don't necessarily think this point needs to be resigned to the trash bin. But, I do not think it supports your position as you think it does.
5) This was your statement: "restriction enzyme sites: P(few but no long stretches|L)/P(few but no long stretches|Z) >10" So...where is it indicated that the size is not the point??? "but no long stretches" seems to suggest size is important. But, then you said "So all 10 synthetic viruses have short maximal segment sizes" which is about the size of the segments! I am not following. You said it is not about size but then proceed to talk about the size.
With regard to this topic, I find a mass of sources with the consensus against you. https://www.biorxiv.org/content/10.1101/2022.10.18.512756v1 argues that restriction sites suggest synthetic origin. https://www.bacpop.org/blog/synthetic/ this is the peer-review comments of the previous citation; it finds the conclusion of the first paper unjustified. https://medicalxpress.com/news/2022-10-scientific-response-sars-cov-.html This is another response to the first citation that finds its results questionable at best. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9937728/ argues that the "endonuclease fingerprint does not indicate a synthetic origin."
Your claim this time is that synthetic corona virus have short maximal segment sizes, and they have 5-8 of them. How many do natural viruses have? How long are the segment sizes of natural viruses? You have not shared enough information for me to grant that you "pick[ed] up another likelihood factor."
Given the papers I cited, I am not going to give you this point.
Again, I reiterate my point, your numbers are either not justified or only support your position when you are performing an inappropriate statistical test.
That's very long and I have some other obligations, so I'll just start with the beginning for now.
1. I absolutely agree that the priors favor Z. I a piece I wrote 2 years ago, I used a factor of 30, acknowledging that was very rough. One of the lest few pandemics came from human error (1977 H1N1) so the priors would actually be higher than 1/30 for L/Z but H1N1 was a restart of an old virus. So it's hard to calibrate because it used t be only old viruses were present in labs but now there are many modified ones. I'm working on a draft now where I give Z 100/1 odds.
2. Jesse says that nothing can be concluded from the raccoon dog data and I agree. The absence of any identified infected intermediate host is unlikely for zoonotic origins but not extremely unlikely. I think 5 of 7 previous new coronavirus disease hosts have been identified, with the 2 unidentified being minor diseases that didn't motivate much follow-up. So this factor favors L but not very strongly.
3.You forgot to square.
4. I didn't actually quantitatively use the synonymous part, just commented on it.
5. Segments are the stretches between the tiny sites, not the sites! The original Bruttel paper used somewhat cherry-picked comparisons in which SC2 stood out as the most extreme of over 1000 sequences. They've toned it down to avoid assuming that the restriction enzymes relevant to SC2 would be the same ones relevant to other viruses, etc. I'm not using the full likelihood ratio that you'd get from their current analysis but just a much milder one consistent with their critics.
Bottom line- I'm on record (https://www.dailykos.com/stories/2021/6/6/2033930/-Zoonosis-or-Lab-Leak) as thinking that the priors heavily favor Z but the location heavily favors L, giving a rough wash. Then the combination of DEFUSE and sequence data, especially re FCS, give more likelihood factors favoring L. You only get to use the priors once, not multiplying them in again each time you update.
No obligations to respond here. This site ain't the best for a proper discussion. I'm sure you and I could come to some consensus on how the posteriors line up if given a chalk board, internet, and enough time.
We have moved away from my contention, though. At this point, July 2023, I'm at 50/50 as to which hypothesis is correct. I think an intentional leak has no plausibility. My contention, and the only position I'll argue for in this thread, is that, given the data available in April 2020, Andersen et alia were perfectly justified in not publishing anything but a zoonotic origin.
Thanks, nice note!
I was ~50-50 2 years ago, as I wrote then. I wasn't eager to change because it seemed important that people take both possibilities seriously in making policy.
I think a lot of evidence has come in since then, making it hard to keep to that position honestly. It's particularly painful to see people I like thoughtlessly repeating empty phrases, letting the likes of Rand Paul sound the like grown-ups in the room. Meanwhile dangerous GOF work is happening in Boston and Rotterdam.
As you say, I bet we could come to some approximate agreement since at least we share the basic framework, as does Nate.