Whenever you look at a map which shows the distribution of Y chromosomal haplogroup R1b you see two areas where the frequency seems very high. First, Western Europe has a very high frequency. Before 2010 it was commonly assumed that R1b was the heritage of late Pleistocene European hunter-gatherers. Around 2010 deeper analysis suggested perhaps that this was not so, and that the deepest divisions in the phylogeny of Eurasian R1b could be found to the east. The high frequency of this haplogroup then may have been an artifact of the Holocene.
Ancient DNA has confirmed this hypothesis. The high frequency of R1b in Western Europe seems to date to the Bronze Age. Though R1b is not found exclusively in Indo-European peoples and existed at low frequencies in Pleistocene Europe, its current ubiquity in Europe seems likely related to demographic turnover between 3 and 5 thousand years ago.
If I had to bet I think R1b, like R1a, originates among the North Eurasian people who mixed with West Eurasians and Amerindians. The Ma’lta boy, for example, seems to have been a basal R.
But notice a secondary mode of R1b in Africa. This is R-V88. The highest frequencies of this Y chromosomal haplogroup are found in Chadic speaking populations. Chadic is a basal group in the Afro-Asiatic language family. A few years ago a paper was published using autosomal DNA on Chad populations and suggested that Eurasian backflow occurred in deep antiquity. From that paper:
We estimate that [autosomal] mixture occurred 4,750–7,200 ya, thus after the Neolithic transition in the Near East…In Chad, we found a Y chromosome lineage (R1b-V88) that we estimate emerged during the same period 5,700–7,300 ya
A new paper, The peopling of the last Green Sahara revealed by high-coverage resequencing of trans-Saharan patrilineages, really gets to the origin of R-V88, with a massive Y data-set. There’s a lot of other Y lineages that are surveyed in this work, but in the supplements, the figure makes it clear that Sardinian R-V88 is basal to star-like African topologies. The implication here is that the African lineages derive from European ones.
The autosomal paper found Chad populations (though the one in question was not Chadic speaking) seem to share drift from Sardinians in particular. Looking at ancient genomes Early European Farmers seem to have been the primary donor population. Additionally, the coalescence of the African lineages seems to date to 5 to 6 thousand years before the present.
Though not definitive, the association of Afro-Asiatic populations with R-V88 is strongly suggestive to me of the possibility that some western Near Eastern Farmers spoke Afro-Asiatic languages.
The results are not inconsistent with the hypothesis that I articulated here: http://dispatchesfromturtleisland.blogspot.com/2017/09/the-source-of-proto-chadic-y-dna-r1b.html in terms of timing and direction of gene flow.
I think it is unlikely that the migrants were from the Near East, because R1 is a more northerly clade, and while western Near Eastern farmers may very well have spoken Afro-Asiatic languages, this is probably a case of language shift by migrants from the Balkans rather than language imposition by the migrants, as one would expect Chadic and Semitic languages to be more similar to each other otherwise.
Weren’t clans matrilineal? Following the Y chromosome would not be representative of tribes and cultures, but of nomads?
Re: specifically Sardinian admixture in Chadic populations, bear in mind that I think from the paper linked this is from f3 stats, and with f3 we probably have to look more at the issues raised by the Guanche adna paper – http://www.cell.com/current-biology/fulltext/S0960-9822(17)31257-5.
Short version, Sardinians peaking f3 shared drift with Guanches, though simple f2 relatedness measures and Fst, ADMIXTURE, PCA etc suggest different closest population (see Gunther’s tweet here – https://twitter.com/trstn_gnthr/status/923580601118097409). Also the case that Sardinians have high f3 shared drift with Levant Neolithic, despite above measures (Fst, ADMIXTURE, PCA, probably f2) showing they are not closest population.
(This is also an issue comparing outgroup f3 stats, to detect shared drift between modern day ME populations and others, e.g. if I recall correctly, even on f3(Outgroup,X,Ashkenazi Jewish) Southern European X inc. Sardinians often seems to coalesce with a random Ashkenazi Jewish sample more than another sample from the same group! Also doesn’t seem to be specific to African outgroups and show up using the chimpanzee as the reference outgroup.).
Strongest stat of the form f3(ADMIXED; X, Y) also doesn’t necessarily (I think) mean that X and Y are the actual admixing populations, just that ADMIXED most strongly violates forming a clade with either X or Y…. e.g. Yamnaya strongly violates forming a clade with Iran_Chalcolithic and EHG… though actual admixing populations were probably not this, IMO, and a model of Iran_Chalcolithic and EHG would probably share less drift with real Yamnaya than CHG+Anatolia_N+Samara_Eneolithic…
Re: R (as in the mutation directly distinguishing R from P) as ultimately North Eurasian, agree probably most likely on balance, though R1 / R2 could have been evolving with Villabruna related populations and Iran_N populations respectively for a long time before the first samples come to our notice (respectively in the Villabruna sample and the Iran_N samples). There’s quite a lot of R1b about in various Mesolithic Villabruna related samples so far…
Strongest stat of the form f3(ADMIXED; X, Y) also doesn’t necessarily (I think) mean that X and Y are the actual admixing populations
yeah. david reich pointed this out to me once.
Weren’t clans matrilineal? Following the Y chromosome would not be representative of tribes and cultures, but of nomads?
the tuareg were. don’t think that’s that relevant in this case.
I used to think that R1b-V88 spread to Africa from the Levantine Neolithic, given its presence in the Near East.
In Lazaridis et al(2016), Natufians could be modeled as a mixture of Basal Eurasians and a population related to WHG, maybe R1b-V88 accompanied this northern admixture to the Levant?
>If I had to bet I think R1b, like R1a, originates among the North Eurasian people who mixed with West Eurasians and Amerindians. The Ma’lta boy, for example, seems to have been a basal R.
Is this why lots of white Americans think they have an ancestor that looks sorta Native American?
The oldest, most basal clades of R-V88 are M18 and V35(found in Corsica/Sardinia/Italy). Yet Schroeder et al (2015) paper ‘Genome-wide ancestry of 17th-century enslaved Africans from the Caribbean’ reports findings of V35 in ancient DNA among some of these non-European admixed African slaves. Africa is still chronically under-sampled relative to Europe. R-V88 in Sardinia I understand is characteristic of populations in more remote inland areas, indicating it’s antiquity. The #1 & #2 peak frequency HLA haplotypes of these same populations are African-derived. Plus there is diversity of African Y haplogroup ‘E’ and ‘A’in Sardinia also. So I suspect R-V88 in Sardinia came from Africa and not vice-versa.
@Chris Davies
I don’t think Schroeder et al say it *is* R1b-V35, but a sister group to R1b-V35, which is what R1b-Y8447 is. Or R1b-V4963 in this new paper, seems to be the same thing.
I assume Sardinians getting some of the highest shared drift with many of those older populations that were heavy in Anatolia or Levant (who were basically sibling populations from a West Eurasian perspective) is just due to not having the really distinct ANE ancestry in Iran or steppe form in appreciable amounts unlike basically every other West Eurasian populations today.