I’ve been recommending the Rootclaim article on the Covid origin question for a while as an example of Bayesian reasoning (with likelihood ratios, priors and posteriors, etc.) that features reasoning so transparent that it seems valuable irrespective of whether it’s correct or not. That is, if your priors on the various Covid origin hypotheses differ, or your interpretation of various pieces of evidence differs, then your conclusions will also differ, but at least you could argue fruitfully with the authors of the piece.
Very interesting page! I think the most Bayesian flaw is not conditioning on earlier evidence in some places. The equivalent of trying to evaluate P(virus with properties A, B, and C) as P(A) x P(B) x P(C), rather than the correct method P(A) x P(B given A) x P(C given A and B).
I’m specifically thinking of how the genetics of cov-2 are not independent of its adaptedness to humans. Once you condition on the unlikely genetics, it shouldn’t be as unlikely that you see transmission in humans, and vice versa.
Substantively, I’m not sure whether coincidences are being cherrypicked here—there are lots of cleavage sites in viruses, and lots of other viruses to compare chunks of genes to. How many similar coincidences should we expect just due to chance? Basically if you update on the coincidences you see but never correct for the number of possible coincidences you could have seen, you’ll overrate how much evidence coincidences give.
I also feel like they put a lot of thought into evidence capable of testing a normal zoonotic origin ( as is proper—https://www.lesswrong.com/posts/rmAbiEKQDpDnZzcRf/positive-bias-look-into-the-dark ), but didn’t put forth similar effort into what goes against the genetic engineering hypothesis. How likely is it that a research lab finds a new bat coronavirus and then before publishing anything about it, decides that it’s the perfect testbed for dramatic gain of function research? This likelihood could be evaluated by checking for prior examples of ambitious research using a microbe before anything about it had been published. This sort of thing is missing.
A hell of a lot of what they say about furin is simply wrong. Insertions and deletions happen ALL THE TIME in nature and while rarer than SNPs are not uncommon, around the cleavage site loop there are other indels in related viruses, the furin site is DECIDEDLY suboptimal for cleavage such that all the fun lineages that are more contagious are optimizing it and creating actual canonical sites rather than something just good enough, it is generated out of frame in in a way that no sane biologist ever would, and the glycosylation pattern nearby weakly suggests it appeared in the context of an actual immune system rather than cell culture.
The alignments they point to arguing for other differences in the cleavage sites in other viruses being due to SNPs rather than insertions are laughable, and from people who I have found to have a poor grasp of the underlying science.
On top of all that when conditioning on something that successfully changes species you will enrich for things that broaden their tropism, which that kind of cleavage site will do, so of course things that are rarer but helpful will pop up more than if you were looking at the space of all possible mutations without conditioning.
How likely is it that a research lab finds a new bat coronavirus and then before publishing anything about it, decides that it’s the perfect testbed for dramatic gain of function research?
In China? We’re talking about a virus based on RNA, which mutates more easily, making it harder to control. China’s government craves control to the point of trying to censor all mentions of Winnie the Pooh, possibly because a loss of control could mean the guillotine.
I’ve been recommending the Rootclaim article on the Covid origin question for a while as an example of Bayesian reasoning (with likelihood ratios, priors and posteriors, etc.) that features reasoning so transparent that it seems valuable irrespective of whether it’s correct or not. That is, if your priors on the various Covid origin hypotheses differ, or your interpretation of various pieces of evidence differs, then your conclusions will also differ, but at least you could argue fruitfully with the authors of the piece.
Very interesting page! I think the most Bayesian flaw is not conditioning on earlier evidence in some places. The equivalent of trying to evaluate P(virus with properties A, B, and C) as P(A) x P(B) x P(C), rather than the correct method P(A) x P(B given A) x P(C given A and B).
I’m specifically thinking of how the genetics of cov-2 are not independent of its adaptedness to humans. Once you condition on the unlikely genetics, it shouldn’t be as unlikely that you see transmission in humans, and vice versa.
Substantively, I’m not sure whether coincidences are being cherrypicked here—there are lots of cleavage sites in viruses, and lots of other viruses to compare chunks of genes to. How many similar coincidences should we expect just due to chance? Basically if you update on the coincidences you see but never correct for the number of possible coincidences you could have seen, you’ll overrate how much evidence coincidences give.
I also feel like they put a lot of thought into evidence capable of testing a normal zoonotic origin ( as is proper—https://www.lesswrong.com/posts/rmAbiEKQDpDnZzcRf/positive-bias-look-into-the-dark ), but didn’t put forth similar effort into what goes against the genetic engineering hypothesis. How likely is it that a research lab finds a new bat coronavirus and then before publishing anything about it, decides that it’s the perfect testbed for dramatic gain of function research? This likelihood could be evaluated by checking for prior examples of ambitious research using a microbe before anything about it had been published. This sort of thing is missing.
A hell of a lot of what they say about furin is simply wrong. Insertions and deletions happen ALL THE TIME in nature and while rarer than SNPs are not uncommon, around the cleavage site loop there are other indels in related viruses, the furin site is DECIDEDLY suboptimal for cleavage such that all the fun lineages that are more contagious are optimizing it and creating actual canonical sites rather than something just good enough, it is generated out of frame in in a way that no sane biologist ever would, and the glycosylation pattern nearby weakly suggests it appeared in the context of an actual immune system rather than cell culture.
The alignments they point to arguing for other differences in the cleavage sites in other viruses being due to SNPs rather than insertions are laughable, and from people who I have found to have a poor grasp of the underlying science.
On top of all that when conditioning on something that successfully changes species you will enrich for things that broaden their tropism, which that kind of cleavage site will do, so of course things that are rarer but helpful will pop up more than if you were looking at the space of all possible mutations without conditioning.
In China? We’re talking about a virus based on RNA, which mutates more easily, making it harder to control. China’s government craves control to the point of trying to censor all mentions of Winnie the Pooh, possibly because a loss of control could mean the guillotine.