View Full Version : mtDNA & Y-DNA E predictions for Plaster et al. Horn Africans by Passa

05-27-2017, 01:27 PM
So, Passa recently came to me with some interesting data he'd managed to compile:

I just wanted to say that I found a way to predict mtDNA haplogroups from Plaster (http://discovery.ucl.ac.uk/1331901/) haplotypes, the predictions look solid enough

Eventually he managed to send the Sheikh both mtDNA and Y-DNA E predictions:

Plaster et al. Y-DNA E subclades predicted by Passa (https://docs.google.com/spreadsheets/d/1hHEJ6z1_QPwYRyPAgYLEb_dt5fRBITrKvI0ezByNvTk/edit?usp=sharing)

Plaster et al. mtDNA Haplogroups predicted by Passa (https://docs.google.com/spreadsheets/d/1XeDxPvM1BfakKyiiYmzcWKw2wZICaPEfUxbzccqdtQE/edit?usp=sharing)

(scroll down to eventually see the frequencies among the various Horn African ethnic groups)

Passa notes the following regarding the accuracy of the results:

By the way, Y-DNA E predictions table is 100% secure when it comes to E-M329 and E-V32, while other sub-clades are definitely less secure
In many cases I couldn't distinguish between V1515+ and M34+, so I put these in the Z830+ category

Well for what it's worth, the predictions were all with a confidence % of >65%
Most of them around 80%

I haven't really reviewed the Y-DNA results all that much yet but I definitely looked through the mtDNA results and created a mtDNA equivalent (https://docs.google.com/spreadsheets/d/1V9GxurCiqtNcJ5XyQ7K3pjjlAMzum-EWYKTm6FuHxHU/pubchart?oid=1540536967&format=interactive) to that old Plaster et al. Y-DNA chart I'd made a while back (https://docs.google.com/spreadsheets/d/1V9GxurCiqtNcJ5XyQ7K3pjjlAMzum-EWYKTm6FuHxHU/pubchart?oid=1059464289&format=interactive) and the results seem sensible for the ethnic groups whose mtDNA results I'm familiar with via other datasets. For example, Somalis turn up as 38-39% mtDNA M+N in Mikkelsen et al. (n=190) and 32-33% mtDNA M+N in a 23andme dataset I compiled on my own (n=75) as well as 40-41% in Boattini et al. once you remove the outlier non-Ethnic Somali results as a commenter at Anthromadness did here (http://anthromadness.blogspot.ae/2015/08/somali-mtdna-frequencies.html) (n=132) so Somalis turning up, for instance, as ~32% mtDNA M+N seems somewhat sensible and close to other datasets. I noticed similar in-line-ness with the prior data for groups like Tigrinyas and Amharas as well. Even in terms of the types of HGs these groups don't show or what have you like the Somalis not turning up with mtDNA X or T or Somalis, Oromos, Tigrinyas and Amharas showing L0, L2 and L3 the most among the non M&N L clades.

Anyway, if these are quite on-point... I gotta say that I'm quite taken aback by how every Sidamic (https://en.wikipedia.org/wiki/Highland_East_Cushitic_languages) speaking group other than the Sidama show overwhelmingly African mtDNA profiles. I assume this is due to admixture with groups like the Omotic speakers in their midst within the SNNP region of Ethiopia (http://i.imgur.com/uJ6iK0A.png). They do show elevated frequencies for L4 and L5 compared to the other ethnic groups (excluding the Gurage who also basically live in the SNNP region but are Ethiopian-Semitic speakers) and those do seem to be notably present among some Omotic speakers (http://anthromadness.blogspot.ae/2016/06/omotic-speakers-uniparental-data.html), I guess. However, I've heard of some auDNA results from a few people of one or two of these Sidamic speaking ethnic groups and they didn't apparently seem much more African than Somalis so I'd caution against assuming they're only ~20% West-Eurasian or something (see here (https://www.ethnologue.com/subgroups/harari-east-gurage) as well).

At any rate, I'll cut my own comments off there and well... Feel free to discuss the overall results for every ethnic group. :)

05-27-2017, 06:39 PM
Interesting how diverse the Eurasian mtDNA is.

05-27-2017, 08:22 PM
I was hoping to finally find an instance of my maternal haplogroup (L3x2b); it's lack of detection from this wide array of populations probably confirms it coalesced outside the horn

05-28-2017, 12:00 PM
If this is accurate, I think it's most revealing for the previously less studied southern Ethiopian groups. Omotic speakers in particular.

We see here that certain Omotic ethnicities are high in M329, the Paleolithic lineage also found in Mota. Others have more M34, which has a YFull TMRCA of 15 kybp, and ties them to Afroasiatic speakers in the Middle East and Egypt. Passa's analysis shows an M34 peak in Genta (65%), followed by Kore (64%), Dorze (52%), Ganjule (52%) and Zayse (51%). Groups high in M329 are Basketo (71%), Ari (67%,), Mejenger (53%). With the exception of Mejenger (Surmic NS), these are all Omotic speakers, which fits rather well with how Ari are particularly close to Mota. It would be interesting to find out how the autosomal affinity of Omotics high in M34 compares to Mota and Cushites. It would be interesting to also compare a M329/M34 ratio between the southern Ethiopians, rather than just comparing the frequencies straight off.

M34 and M329 may also help trace Omotic admixture in southern Ethiopian Cushitic speakers. Konso, Gewada, Dirasha, Busa and Burji all show a combined M34+M329 frequency of >40%, as high as 67% in Busa.

M281 shows a surprisingly wide distribution at low levels here, from south to north. The southern groups are largely unknown (although M281 was reported in 4% of the Maale, highest so far), but otherwise M281 is usually scarce in previous data from non-Omotic groups, yet here it shows in 4-6% of Amhara/Tigray. It may partially represent some other lineage(s).

The Hadiya (Sidamic) show rather high >50% unspecified Z830, so there's some potential for finding a new subclade there, or a higher frequency of one of the less common M35/Z830/V1515 clades.

Focused on the Y because the results seem more specific. Not because moms don't matter. :)

06-28-2017, 12:52 PM
Made maps for most Y-DNA and mtDNA lineages: https://www.dropbox.com/sh/00whitx0iyglua6/AAD8X733iKrRM2t_hO5d3eHMa?dl=0