January 2011
Comments by Argiedude:
A recent study (Naidoo, 2010) looked at the y-dna of South African whites, Bantus, and Khoisan. Here's a link to the pdf: http://www.investigativegenetics.com/content/pdf/2041-2223-1-6.pdf
[To view the actual data used for the report, got to http://mbe.oxfordjournals.org/content/early/2010/11/25/molbev.msq312/suppl/DC1 and click "Supplemental Table S1 to S3".]
The study tested for both E1a* (M33) and E1a1 (M44). It found only a single sample, that belonged to E1a1 (M44). And it was in a South African white. Interestingly, there are some samples from Afrikaners in yhrd, one of which has all the telltale signs of belonging to E1a1, such as 391=9, 392=12, 390=22, 19=16, 389a=12. There's no haplotype data for the Naidoo sample.
One case of E1a1 in Afrikaners is wild luck, but 2 samples in 2 studies... this is very weird, it's extremely unlikely to be a wild coincidence statistically. Afrikaners are from northwest Europe, and in northwest Europe E1a1 is present at a rate of 1 in 10,000 samples. These 2 studies combined (yhrd + Naidoo) have 265 samples from Afrikaners/whites, and there are 2 samples of E1a1. It's a little too much for a wild coincidence.
Now obviously one would think this must be a case of historic South African Bantu/Khoisan ancestry in Afrikaners. But the Naidoo study confirms yet again that neither E1a* nor E1a1 exists at all in the southern half of Africa. At this point, thanks to the Naidoo study and also the other big recent study of southern Africa by de Filippo, which I'll comment in another thread (that's why I said I had 2 things to talk about), the accumulated southern African y-dna samples must be around 3,000, and there isn't a single sample amongst all of them of either E1a* or E1a1.
There's no way to justify this appearance of E1a1 as the result of ancestry from an outside group, wether it be other Africans, Jews, south Europeans... nothing fits. I don't have any explanation, except that it must have been genetic drift increasing a pre-existent lineage to its current levels, but that too would require a remarkably far- fetched coincidence: the genetic drift, acting on any random sample found in the original northwest population, just had to fall upon one of the most exceptionally rare lineages to be found amongst northwest Europeans (1 in 10,000 samples).
So this to me is a true mystery. We can assign E1a1 to some 3clusters: West Asian, Jewish [or just Ashkenazi?], and Sahelian West African. The Jewish cluster can be discarded, at least for the yhrd E1a1, on account that it has 385a/b=16/16, while the Jewish cluster has the distinctive values 385a/b=15/17.
There was a recent huge study of African y-dna (de Filippo, 2010), which included more than 30 E1a samples. The study tested M33 (E1a), but not M44 (E1a1).
Judging from the haplotypes, it would seem all the E1a are E1a* except one probable E1a1: HGDP01284, from Senegal.
The Senegal E1a1, apart from the signature 391=9 + 392=12 of all E1a1, has 385a/b=16/16, which is one of the features of the West African cluster of E1a1.
I also think the following corrections should be made:
BF199 is E1a but should be E1b1a
BF157 is E1b1a but should be E1a
BF202 is E1b1a but should be E1a
BF082 is E* but should be E1a
DK016 is E1a but should be E*
So there you have it, an unusually large collection of SNP verified E1a samples, perhaps the biggest in any study yet. ,