Fascinating: reality never stops surprising us. All the basal phylogeny of human Y-DNA has been revised quite radically.
Now super-haplogroup BT, Y(xA) or (my personal favored name) B'CDEF is just a branch, a sublineage of A. And not just of A but of a fraction of a fraction of it:
To the left we have the new basal phylogeny of the human Y-DNA. Please ignore the "time" axis. I'll get to that later.
We can well ignore the proposed nomenclature and begin talking of A as the pan-human lineage or "Y-DNA Adam". This primeval lineage split then into A1b (found at 8.3% among Bakola Pygmies and 1.5% among Mozabites) and A1a-T (or Y(xA1b) or A1a'2'B'CDEF, using mtDNA's "superglue" nomenclature's fix-it-all method), which includes all the rest of the World's population.
Don't worry too much about the nomenclature because it will obviously cause an upheaval in the naming conventions of at least A. All these names are surely just provisional.
Anyhow, this second lineage is yet split between A1a and A2'3'B'CDEF (A2-T in the graph). A1a is found mostly in Niger: among the Fulbe (14.9%) and their Tuareg neighbors (4.5%), as well as among Middle Atlas Berbers of NE Morocco (2.9%).
Importantly Sudan and other large areas of Africa SE Africa, much of Middle Africa, were not sampled. But I'd say that this distribution is suggestive of an origin near Chad Lake in relation with the expansion
of mtDNA L1
. But I'd wait until we have results from Sudan because Sudan, notably the South, hosts huge Y-DNA diversity and a lot of Y-DNA A, so it cannot be ignored so easily.
as quick reference, we can see that Y-DNA A1a has also been reported among several peoples of Guinea-Bissau (2.8-5.1%), notably the Mankanha speakers (7.8%). Also 5% of Mandinka from Senegal and Gambia, 2% of Dogon from Mali, 3% of Moroccan Berbers and 2% of unspecified Malians. Basically we can say that A1a is somewhat common (always at low frequencies) through West Africa.
However no further data on the crucially basal A1b lineage is found in this article at least. Anyone?
Besides, the next division is between A2, A3 and the remainder: B'CDEF (or Y(xA) or B-T). A2 is exclusively a Khoisan lineage, while A3 is split between Khoisans and the peoples of the Upper Nile (Sudan, Ethiopia...), having greater basal diversity in this last region (A3a and A3b2 are both Sudanese/Ethiopian, while only A3b1 is a Khoisan lineage).
In any case, the pattern found now in Y-DNA seems to suggest a flow from West or Central Africa to East Africa and then to Southern Africa (Khoisan peoples). This is somewhat contradictory with the pattern revealed by mtDNA (suggesting an origin in East Africa around where the oldest fossil H. sapiens are known to have existed, by the Omo river) or the pattern suggested by some autosomal diversity measures performed recently
that proposed Southern or Central-SW Africa instead.
Typically they forget to (or could not) take samples in crucial areas like Angola (other than Khoisans), Mozambique, Tanzania (not even the Haza and Sandawe) or Sudan:
|Fig. S1 (see: supp. materials). |
Red: A1b, green: A1a, black asterisks locations of A1a from other papers.
The technical differences between the old phylogeny and the new one are explained in detail and best perceived in fig. 2 (below). Besides of finding a large list of new SNPs, Cruciani et al. suggest that M91 is an unstable (and hence unreliable) location, while P108 and P97 are stable but were interpreted wrongly when first described, being the polymorphisms the ancestral ones, i.e. those identical to what is found among Chimpanzees.
The authors propose a time-line for the human Y-DNA that makes no sense. Notably CDEF (CT) is proposed to have only 39 Ka, being more recent even that Aurignacian and several findings of Eurasian remains of H. sapiens (not even considering Skhul and Qahfez). It is also contradictory with so many proposals of Y-DNA timelines that suggest a CDEF age of c. 70 or 80 Ka. (Karafet 2008 for example), which I still consider a bit too recent - mind you.
At the moment there two possible archaeologically consistent scenarios for the Out of Africa migration, one would have happened c. 90 Ka ago, reaching South Asia by c. 80 Ka (Petraglia 2010
). The other one, proposed by Armitage
this year (discussed here
, see also here
), would have the coastal migration by South Arabia happening c. 125 Ka ago and could have reached South Asia soon after. There could even have been two successive migrations.
As Y-DNA CDEF is central to the Out of Africa migration, it must have coalesced before this one happened. Unless one would argue for a back-migration of DE into Africa, what I find hard to sustain. Even if you'd do that, CDEF would have need to be consolidated c. 80 Ka ago at the latest.
That is double than suggested in this paper. It could be triple or even more. That would make "Y-DNA Adam" (everything else equal) at least as old as 285 Ka, surely more.
Of course, I am not believing a word on Y-DNA age estimates, because the hunches are so wildly different and so poorly argued (not to mention that no scientific proof was ever provided supporting such statistical methods) that it's a total waste of time.
The phylogeny however stands and is an invaluable piece of information.
Important update (May 27): Other possible populations with A1b
believes that he had identified A1b lineages in a number of populations (based on other studies and using mostly STR haplotype sequences). He included this spreadsheet
(should be available for a year) where the likely A1b individuals were identified by him as "A4". See the commentaries for further details.
The populations with possible A1b are therefore expanded from just two (Bakola Pygmies and Mozabites) to several, with a peak of frequency in Southern Ghana. Also in Gabon, SE Nigeria and Cameroon, and as isolated individuals in Zambia, Ivory Coast, Burkina Faso (and then, of course, across the Atlantic in America).
Update (May 2 2012):
Marnie reports at her blog
that A1b (the earliest splitting-branch) is now being renamed A0
and that it has been detected in an African-American man previously matched to modern Cameroonians from near Bouea.
Also the Family Tree Y-DNA A project
has detected two branches of A0 (former A1b), although one is still a private lineage
(too rare to be accepted as a distinct haplogroup