[tripletao]

> That argument points out that genetic drift is well characterized and the presence of two closely related SARS-CoV2 variants cultured from the wet market is extremely strong evidence that is where the virus initially appeared.

I assume you're referring to Pekar et al. here? The two lineages are literally just two SNPs apart, so it's near-impossible to distinguish whether they arose from two separate introductions, or just from two super-spreading events after cryptic evolution in humans from a single earlier introduction. Pekar builds an epidemiological model that purports to find that evolution in humans is p ~ 0.5% unlikely; but that result is highly sensitive to the assumptions in that model, most notably their choice of a scale-free infection network (and thus power-law distribution of number of other people each patient infects). Robustness to that infection network isn't studied.

The author of this endonuclease fingerprint preprint also has a preprint on Pekar's model,

https://www.biorxiv.org/content/10.1101/2022.10.10.511625v1

Note that I'm criticizing Pekar here, not endorsing the endonuclease preprint. I don't have a great sense of the correct Bonferroni correction (to borrow Prof. Balloux's framing) to apply to the latter's probabilities.