@anon46279830 would Amos join us here in a scholar thread to explain his case to several well qualified scientists who really want to understand his point?
@swamidass
You read my mind. If I had some credentials from five or six of you that would be willing to join in I believe I could persuade him. He is across the pond of course so it would have to be over the course of a few days. Iād prefer to get out of the middle and let people in the field scrutinize each other directly. Once it gets technical and not conceptual, Iām not gonna be able to hold up.
I have the sheet on Josh. So who else would be willing and what might I tell him about your professional background?
Iām sure @glipsnort and @John_Harshman would join in, as would I. If we are lucky, so would @Joe_Felsenstein. If that is what he wants, have him email me a brief statement laying out his case, even if it is mainly just links to the key articles. I will use it to start an office hours and we will wait for him to join.
Itās made quite clear in the next sentence:
We therefore conclude that around 3% of SNVs are likely the result of MNM events.
And one of the preceding sentences (my emphasis):
Assuming that mutations are independent we would expect 0.03% (95% c.i.: 0.00%ā0.07%) of mutations within this distance of each other.
That is one which should definitely be explored. I had seen another paper from a year or two ago which had come to much the same conclusion. If the diversity in Sub-Saharan Africans is largely an artifact of another introgression- from something so close to the rest of us that it hasnāt even been caught until now, say an Archaic Homo Sapiens or even just a strain of Homo Sapiens that otherwise died out - then it could serve to undermine Dr. Amosā hypothesis. You could claim his 20,000 pound prize! But to do so will I think leave the door open to an even bigger sea change in the world of human origins. If most of human diversity is an artifact of introgression then we came along as a distinct group way more recently than is otherwise considered or mutation rates are way slower than otherwise considered. Or both.
It is also possible that what your link implies is true but that Amos is still right. Especially if part of the answer is slower mutation rates.
Is that right @glipsnort and @John_Harshman and @Joe_Felsenstein ? It seems like glipsnort in particular is busy these days. If you guys say you will take time to engage, I will try to recruit him. I would not want to bring him here and then there be no engagement is all.
Sure, I would participate, though Iām by no means an expert on this subject. Professionally, I try to ignore within-species variation as much as possible and hope (in vain) that neither introgression nor ILS ever happens.
Incidentally, the paper @glipsnort wants us to read is behind a paywall.
2 Likes
Yeah I canāt get to the @glipsnort link but here is a link to a prior study which comes to very similar conclusions. The math is beyond me, but I get the bottom line. https://www.biorxiv.org/content/10.1101/489401v1
Not interested in participating. Too busy with too many other projects, this one involves too much consideration of details of the mutation process and of the detection of Neanderthal DNA.
1 Like
Iām unlikely to have time during the next month. After that I could probably look at it.
1 Like
I havenāt seen anyone suggesting that most sub-Saharan diversity results from archaic admixture, just that the admixture occurred and is detectable. The reason I recommended that particular paper (sorry about the paywall) was something more specific, though. In that study, the authors use a haplotype-based method to look for evidence of archaic admixture; theyāre looking for long(ish) pieces of chromosomes that are highly diverged between different copies. To test their method, they apply it to samples from Melanesian individuals and see whether they can, without referencing the Denisovan genome, recover the signal of Denisovan introgression. The did so. Not perfectly, but quite well: they found a strong signal of introgression, and the introgressed elements corresponded very well with what could be found when the Denisovan genome was used for comparison.
To me, this is strong evidence that haplotype-based methods like this work very well for detecting introgression.
3 Likes
And if the high diversity is contained within discrete fractions of chromosomes, that would seem to argue against Amosās explanation and in favor of the standard one.
The most diversity in the sense that the most diverse genes were not from mutations in the same species diversifying over time. So if the genes with the oldest MRCA were introgressed, then it doesnāt count as far as when the main stock of humanity split. It would give a false appearance of age in the split.
But regarding your point about looking for matches in long segments, if mutations are linked as he claims then it seems like youād still have to be careful about false positives.
Sure, introgression increases the mean time to MRCA (or equivalently, it increases the effective population size), but not by a lot. Thatās because the mean tMRCA even without introgression is already older than the likely split time from any population that would be the source of the introgressed DNA. Suppose 5% of the genome in a population stems from introgression from a population that split 700,000 years ago (rather like Denisovan DNA in Melanesians). If you compare two modern chromosomes, more than 90% of the time youāre comparing non-introgressed sequence ā and yet the mean tMRCA for that comparison is still ~1 million years.
One should always be careful about false positives. My guess is that linked mutations are a small enough effect to cause only modest problems. But thatās why I pointed to the African introgression paper ā Iād much rather see a successful test of a method on a positive control than try to calculate error rates in a model.
PS- if you think you are going to get a window next month, I request a heads up. If you and a couple of others would be willing to dialogue I will try to recruit Dr. Amos to come here and sell it.
I already told you the process. Have him email me.