Wednesday, October 5, 2016

Sudanese Arab and Nubian mtDNA is mostly non-Eurasian?

I was looking back at some old data like from Hisham Yousif Hassan Mohamed's 2009 thesis and there's something intriguing about Sudanese Arab and Nubian mtDNA:

There aren't too many samples but there's definitely a pattern going on with the Nubians and the Arakien, Gaalien and Meseria (the three groups used to create the "Arab" grouping above); most of their mtDNA is non-M&N. The overwhelming majority of their maternal lineages look to be African rather than West Eurasian.

This is surprising because it's notably different from how things are with Horn African Cushitic and Ethiopian Semitic speakers as well as Bejas who are a northern Sudanese population like Nubians and Sudanese Arabs (Bejas are Cushitic speakers though):

Somalis, Amharas, Oromos, Tigrinyas, Beta Israels and Bejas all have a more balanced mtDNA profile. 40-60, 50-50, 50-50, 60-40, 50-50 and 30-70 are the types of frequency ratios you see in terms of Eurasian (West Eurasian derived) and African (non-M&N) lineages.

Now, the thesis sample sizes arguably aren't large, to be honest, but it's still notable that Sudanese Arabs and Nubians have very low frequencies for M&N lineages even when the sample sizes are comparable to those of the Horn African groups and Bejas. It's also notable that this pattern holds in all of the groups (the Nubians and the three Arab groups).

Despite this, they actually have high frequencies for West Eurasian Y-DNA lineages (i.e. J1):

Fifty to seventy percent or more of their Y-DNA lineages are of clearly Eurasian (West Eurasian derived) origins. [note] A friend once brought up the idea that this implies much of their West Eurasian ancestry since after the Sudanese Neolithic (something Cushitic speakers from the Horn of Africa seemingly descend from) is probably "male mediated" as in what largely happened is that males carrying substantive West Eurasian ancestry intermixed with local Northeast African women and brought lineages like Y-DNA J & I in abundance.

I do wonder if that's seriously the case...

This would be unlike the case in the Horn where there's about 5-30% Y-DNA J (mainly J1) among Cushitic and Ethiopian Semitic speakers alongside 5-35% Y-DNA T as well. [note]

The Horn looks somewhat more "balanced" in terms of Y-DNA & mtDNA lineages. There's a substantial amount of both blatantly West Eurasian and more local African lineages on both a maternal and paternal level but things look quite skewed in the case of Nubians and Sudanese Arabs who look to have very low M&N lineage frequencies mtDNA wise but very high frequencies for lineages descended from F in terms of Y-DNA.

Should be interesting to see more and more mtDNA & Y-DNA data from Nubians and Arabic speaking groups in Sudan to see how much this pattern holds in the future. Getting mtDNA results from different Arab groups is also key as I've only really seen results for three groups at this point.

Tuesday, October 4, 2016

Southeast Africans and Chadians from Triska et al. 2015

Some other interesting things to note about this paper's new samples are the Southeast African and Chadian samples.

Some simple observations:

  • The Samburu samples look to be highly Horn African-shifted. Their African and West Eurasian ancestry levels seem only a bit lower than those of the Somalis and Oromos. I suppose this implies that they have quite a bit of admixture from East/South Cushitic speakers. Something people have assumed long before genome sequencing, as far as I know... Their mtDNA lineages also implied such admixture, last I checked, so this makes some sense. It's just that the level of this admixture that's a bit surprising.
  • The Turkana samples seem less Horn African-shifted than them but also seem to clearly have substantive admixture from Horn African Cushitic speakers who were probably quite genomically similar to modern Somalis. Again, this was somewhat expected and Tishkoff et al. 2009, albeit with less high-resolution data, ultimately indicated that Turkanas had such admixture anyway. Both them and the Samburus, despite being Nilo-Saharan speakers, are clearly mixtures between Nilotic speaking, Bantu speaking and Cushitic speaking Southeast Africans.
  • The Daza Chadians are an intriguing case too. Despite being Nilo-Saharan speakers; they're clearly rich in West Eurasian ancestry. They're about as shifted toward West Eurasia & North Africa as Fulanis are (pulling about as eastward as they do in the PCA). The main difference, however, is that the Fulanis clearly seem like more of a mixture between West-Central Africans and Northwestern Africans whilst the Daza look to, mostly, be a mixture between the East African cluster and West Eurasians. [note]
  • The Kanembu essentially just look like more West-Central African shifted versions of the Daza, understandably showing less of a pull toward the West Eurasians and North Africans as a result.

What I find most interesting about all of this are the Chadians, quite frankly. What's quite intriguing about their West Eurasian ancestry is that it, based on the ADMIXTURE results above, looks to be mainly derived from Northwestern African-related people (Mozabite-like peoples) as its mainly derived from the light blue cluster. This means they're basically some sort of East African version of the Fulani samples, especially in the Daza samples' case. They seem to lack the more Bedouin and Druze related affinities the Oromo, Somali, Nubian and Sudanese Arab groups are showing.

Although, they also differ from the Fulani samples in that they mostly don't show ancestry from the grey Sardinian peaking cluster. Their West Eurasian elements are wholly swallowed up by the light-blue Mozabite peaking cluster. I don't know what significance that might have but it's worth taking note of.

I really don't know much about Chad's history but I wonder if they used to be Chadic speakers? I'd be interested in seeing what sort of R1b-V88 frequencies these two populations have if anyone's seen Y-DNA data on them... Baqara Arabs (بقرة العرب) like the Messiria tribe seem to have notable R1b-V88 frequencies (see here) which, to me, implies some of their ancestors were Chadic speakers (like the Hausa) before their Arabization.

A Baqara Arab man and child riding cattle

Anyway, I'll leave any further theorizing and observations to future posts that will hopefully be based on new PCAs and perhaps also ADMIXTURE runs where these new samples will be run alongside the various other HSS samples out there.