From TPM Reader MN …
First of all, for credibility’s sake, I am a computational biology postdoc at [*******]. I’ve done some research on the SARS-CoV-2 genome but it hasn’t been my main focus the way it has for many people. Nonetheless, I’m acquainted with at least the discussion of genomic mutations and evolution, although the nitty gritty web lab virology is not my area.
Second, a couple of points on the previous takes. 1) I was delighted to see Harry Frankfurt’s classic work cited, it’s an all time great manuscript. 2) The WHO report had a lot of good information but it was also comical, particularly the repeated non-sequiturs addressing the idea that the virus was imported to Wuhan on the surface of frozen food products. Someone with some influence *really* wanted that possibility to be explored and you could almost feel the writer rolling their eyes. 3) I’m not wild about the Garry write-up or the TWIV discussions that AJ cited. Several of Garry’s points are good but the suggestion that all of it is utterly inconsistent with a lab leak is a leap, particularly the presence of two lineages at two different markets. Sure all else equal those are consistent with a spillover but he suggests that a lab leak would require that they diverged while still in the lab, which is simply not true. Those lineages were different by two nucleotides, which could easily accrue during a few weeks of undetected circulation.
To be clear, I still lean away from the lab leak hypothesis but I have found it intriguing because what’s been different about this round of speculation is that it’s based on something approaching a theory of how it actually happened. Actually Garry discusses it in that write-up, but essentially the thought is that the backbone was this RaTG13 genome that had been mutated via serial passage to something like 96% similarity with the original copy, then the RBD of the spike protein from the pangolin virus was spliced in, then somehow it acquired the cleavage site and it was off to the races. The first of those two seem at least plausible to me, although someone with intimate knowledge of virology and how mutations accrue in those two lab processes might be able to point out details that make that unlikely. The third step seems unlikely to have been by design since the mutation is out of frame, but it’s not impossible. But on that note I want to make two points which are the actual reason I’m writing.
First, one good way to proceed would be to talk to lab members and researchers to understand whether anything like that was done and how frequently the genomes of various viral specimens were sequenced (since their databases have ostensibly already been searched). So it’s definitely a bit annoying that additional cooperation seems to have been ruled out. But second, if this is how it happened then my guess is that there are breadcrumbs somewhere in the various sequencing datasets that the Shi lab and others at WIV have deposited over the years. Digging into those is time consuming and will require a lot of bespoke analysis, but the data are usually publicly available. I would explore this myself if there were more time in the day. A related point is that this sort of thing is a relative niche, so it’s not really subject to the efficient market hypothesis and as a corollary it’s not totally true that this round of discussion is not based on any new information. It’s taken some time for enough dots to be connected that even this half-baked theory has come together, and it’s quite possible we haven’t heard the last word on this.