I presume that the wide clusters called “Asians” also faced the same kind of cold conditions, at least, because their average temperature is similar to the main European group and they instead show divergence levels similar to Tropical Africans.

It’s more like, yes, the GG original allele has evolved out of Africa towards AA but not clearly because of cold but for whichever other reasons, maybe just drift or a very weak selective pressure that would cause a very weak selection signal like the one we see. Asians being less derived in general (because Asia is closer Africa in pgylogenetic distances, not in mere miles) show a lower frequency of derived alleles, while some of the populations of the extremes, probably because founder effects, which may or not have been driven by selection, show a high frequency of derived alleles.

Would it be related to cold, I’d expect Amerindians to be the more derived, because they lived in Siberia and Alaska for a time – and that’s one of the coldest places of Earth, in the Ice Age too. But the few Amerindian samples shown in fig. 4 are congruent with Asian (or sometimes Oceanian) ones and spread around.

That Europeans and Oceanians are representative of adaptation to cold in opposition to Asians and Amerindians is hard to swallow, sincerely.

A possible explanation, although I agree the whole thing is ‘Kind of messy’: it’s usually accepted in this part of the world that Oceanians, especially Polynesians, have been subject to selection as they moved across the ocean. Cold nights during such migration has given rise to the Polynesians’ more robust body form than would be expected for populations living at a comparable latitude on land. Slender bodies were more prone to hypothermia and so tended to be selected against.

Regarding the general thrust of the article, ‘that mtDNA diversity is lower in populations residing in cooler climates’, I would suspect that’s as much a product of periodic extreme selection, or even partial extinction, in such environments rather than being a product of cell metabolism associated with particular mtDNA haplotypes.

First thing that struck me is that the graphs are “corrected for distance from Africa”. However we know by the mtDNA phylogeny that distance from Africa means little in the actual history of Eurasian expansion, that India was first, then East Asia, then Australasia and only then the areas closest to Africa: West Eurasia.

But I still could not make up my mind. I was sort of thinking: to demonstrate this hypothesis you need to correct not for metric distance from Africa but for phylogenetic distance from East Africa probably.

Kind of messy, I know.

But then I look at figure 4 and what do I see? A median that does not correspond at all with any of the clusters. In fact I see clusters happily dancing around the median at whim, organized not by temperature or whatever but by ethno-geographic clusters, regardless of temperature. Most Asians (missing regional distinctions here) well below their theoretical position, West Eurasians above and, notably, Oceanians happily at the same level as Europeans, in spite of having a temperature almost identical to that of Tropical Africa.

This pattern, largely independent of temperature, is repeated for both loci.

So what does this figure 4 says in fact: that most Asians (but not all) cluster best with most Tropical Africans for these two loci, in spite of living at climates almost 20º colder. This clustering is not full and in fact Asians are intermediate, transitional.

But Oceanians and Europeans cluster even more closely in spite of distance and abyssal temperature difference that reaches up to 40º in some cases.

The main Asian and European clusters have the same temperature and are instead completely different for both polymorphisms.

So false positive, sorry.