Citation: Recent evolution of the mutation rate and spectrum in Europeans, Kelley Harris doi: 10.1101/010314
The above figure is from a preprint, Recent evolution of the mutation rate and spectrum in Europeans, which reports very peculiar results from the 1000 Genomes data. I actually got a preview of the topline finding about a year and a half ago at a Bay Area Population Genomics meeting, but many of the details are new to me. As noted in the abstract the “private European variation is enriched for the transition 5’-TCC-3’→5’-TTC-3’.” The implication here is that different populations mutate differently. The preprint puts this in the broader context of the fact that for a while now there have been conflicts between different rates of mutation inferred from pedigree and whole genome sequencing, and phylogenetic models of divergence of species. At this point the technical details need not concern. Rather, let’s just add that the recent ancient Siberian genome paper confirmed this discrepancy, and strongly supports the contention within this preprint that the mutational rates across the ape lineages are likely to have varied, questioning the validity of an invariant molecular clock.
Frankly I trust Harris to be right about the pattern she sees here. She’s been looking at this data for a few years, so if there was any statistical artifact here I am confident either she or her advisers or colleagues would have caught it. But there are some issues with the attempt to integrate these results about differences in mutational spectrum with population history. Some of these are pointed out in the comments at bioRxiv. Aside from the simple semantic conflation of Early European Farmer (EEF) for Eastern European Farmer, the attempt to suggest that reduced enrichment in northern Europe is a function of Ancestral North Eurasian (ANE) admixture is made less persuasive by pointing to the case of Finns, who are known to have a secondary East Asian admixture which arrived from Siberia more recently. I think that this is not a problem when you see another issue “Because East Asians share a more recent common ancestor with ANE than with west Eurasians.” I do not think this is the dominant view. Rather, ANE and West Eurasians are best modeled as a distinct clade with deep common ancestry as against East Eurasians. See figure 3 of Lazaridis et al. The confusion here matters because the thesis being presented seems to be that ANE lacked the enrichment of a particular mutational class, as modern East Asians do. This is a warranted conjecture if the two formed a clade with West Eurasians as an outgroup, but this is just not the case.
Which brings us to when and why the ancestors of Europeans began to exhibit this particular mutational pattern. As it happens the results show that even without the Finnish sample there is a north-south gradient of enrichment toward the latter. This would support a model where ANE admixture resulted a decrease from an originally higher proportion. That would mean then that the change occurred probably in the interval of 20 to 30 thousand years ago, when we presume the ancestors of West Eurasians and Ancestral North Eurasians diverged. But this is not the only option. One element of EFF ancestry is Basal Eurasian, which happens to be a group which is equally distant from West Eurasians, ANE, and East Eurasians. In other words, Basal Eurasians possibly diverged from all these populations before the primary Out of Africa event ~60 thousand years ago. If the mutational spectrum deviation derives from Basal Eurasians then the gradient could be a function of reduced EEF ancestry in Europe as one goes north. Looking at the difference between the Finnish and Italian samples I do not think this is the case, the variation is too small. The EEF fraction varies a great deal in Europe. So the ANE dilution model actually does seem more plausible.
But there’s a final element to be explored. Why is there in enrichment in the first place? It turns out that this sort of mutation is very common in melanomas. In particular of interest to me: “Folate deficiency is known to cause DNA damage including uracil
misincorporation and double-strand breaks, leading in some cases to birth defects and reduced male fertility.” Folate deficiency can occur when light skinned individuals are exposed to sunlight. It strikes me that the higher mutational load for these particular transitions in Southern Europeans as opposed to Northern Europeans could simply be a function of the fact that they are in sunnier climates. We know that Europeans have become much lighter skinned very recently, so the range of mutations we are seeing may be due to very recent factors. No one knows concretely why Europeans became very light skinned very recently, but these mutations may simply be a side effect of this phenotypic change, which was driven by powerful selective forces.
Citation: Recent evolution of the mutation rate and spectrum in Europeans, Kelley Harris, http://dx.doi.org/10.1101/010314
Addendum: It would have been nice of the 1000 Genomes had at least one Middle Eastern population.