4 ms·

Then let me offer two informal "substantial" arguments. It's too long to question every point. The author argued in section "weirdness 1" that the phylogentic

by daxuak 5y ago

Then let me offer two informal "substantial" arguments. It's too long to question every point. The author argued in section "weirdness 1" that the phylogentic tree implies the omicron strain is directly related to an ancestor genome from Spring 2020. I have the following concerns: the tree is based on https://nextstrain.org/ https://nextstrain.org/ , a database initially introduced in 2018 [1]. This is a great data source. However, for COVID, nextstrain naturally bias towards the more viable biosamples. In other words, COVID strains in wild animals such are going to be underrepresented, if represented at all. And if so, the large sequence divergence is surprising but not a direct indication of lab leak. The author can supply with a rough overview of sequencing sources of nextstrain to counter this. Or better, it would be good if the author can compare with, or cite existing literatures, phylogenetic trees of other epidemics and demonstrate that the sequence divergence is much larger than what we would expect from wild animal reservoir. "weirdness 2" has a stand, but it should be accompanied by silent/functional mutation ratios in other COVID strains since synonomous mutation could be not entirely silent if translation efficiency is affected (correct me if I'm wrong). Briefly on one other point, "weirdly it's exactly mice, not other animals": we use mice as lab animals for a reason. They are somewhat similar to human in terms of physiology. Therefore we can expect that they could be more prone to COVID infection, aren't we? editted to add one more sentence: from a peer review aspect, this blog is beyond terrible and I'd probably have 10 major conerns listed if it comes in as a PR. [1] https://academic.oup.com/bioinformatics/article/34/23/4121/5001388 https://academic.oup.com/bioinformatics/article/34/23/4121/5...

howaboutnope 5y ago
> correct me if I'm wrong I don't even remotely have the expertise for that, if applicable others would have to do that. But thanks a lot for elaborating!
- daxuak 5y ago
  Oh I genuinely mean that and nothing else. I'm remotely familiar with the topic, but I do not work with protein translation and viruses at all. My question about this silent/functional ratio argument could be completely wrong or in the wrong direction. That original argument is interesting and is a surprising obervation to me, and I would love to read a well-written piece on that. Just keep in mind that "statistically unlikely" doesn't immediately mean "it's unlikely", you might have used the wrong presumption (in this case, silent mutation being not really silent).
daxuak 5y ago
Adding one more clarification in case someone reads this text wall: The "requests" I made was rather harsh, e.g. it might be impossible to check nextstrains's sequencing sources, if this is a real manuscript revision. I honestly probably won't question as harsh if it's a writting for a different topic. However, given that the author is trying to illustrate a not-very-likely scenario, which being, not only that there's a major lab leak, but also someone has engineered a not yet fully understood non-trivial virus genome to alter its function as desired (note that COVID doesn't have a well established animal model. It would be hard to test a engineered virus even if it exists. In this case, would you assume that it's continuous lab "leak" and only the omicron strain was a success? And if yes, how could all other strains went undetected at all in the past 2 years?), I think scrutiny is justified and should be expected. It is against several odds and you really need some strong evidences to rule out other possibilities.

by daxuak 5y ago

howaboutnope 5y ago

> correct me if I'm wrong I don't even remotely have the expertise for that, if applicable others would have to do that. But thanks a lot for elaborating!

daxuak 5y ago
Oh I genuinely mean that and nothing else. I'm remotely familiar with the topic, but I do not work with protein translation and viruses at all. My question about this silent/functional ratio argument could be completely wrong or in the wrong direction. That original argument is interesting and is a surprising obervation to me, and I would love to read a well-written piece on that. Just keep in mind that "statistically unlikely" doesn't immediately mean "it's unlikely", you might have used the wrong presumption (in this case, silent mutation being not really silent).

daxuak 5y ago

Adding one more clarification in case someone reads this text wall: The "requests" I made was rather harsh, e.g. it might be impossible to check nextstrains's sequencing sources, if this is a real manuscript revision. I honestly probably won't question as harsh if it's a writting for a different topic. However, given that the author is trying to illustrate a not-very-likely scenario, which being, not only that there's a major lab leak, but also someone has engineered a not yet fully understood non-trivial virus genome to alter its function as desired (note that COVID doesn't have a well established animal model. It would be hard to test a engineered virus even if it exists. In this case, would you assume that it's continuous lab "leak" and only the omicron strain was a success? And if yes, how could all other strains went undetected at all in the past 2 years?), I think scrutiny is justified and should be expected. It is against several odds and you really need some strong evidences to rule out other possibilities.