I’m just suspicious about the 60/40 average you came up with, 58/42, and 62/38 means there’s a 4% variation in SSA admixture, might not seem a large difference but that could mean those on either end are distant to each other as a Englishman and a Dutchman!
What? The population results move together. So if the average becomes 62-38 then the variance among Somlis is the usual 2% of
38-40% for Eurasian ancestry depending on the person. If the average becomes 42-58 then the population variation is now
42-44%. Anyway, a 4% variance between the methodologies is not wild at all. It's if anything quite remarkable that two totally different methods (Pickrell's and the PCA positions) come so close in estimates (basically 40/60).
Also curious to know amongst whom does SSA /WE admixture peak?
Among Somalis? No one, really. All Somalis are basically the same as far as I've seen. Very homogenous ethnic group. Only outliers with recent outside admixture (Oromo, Habesha, Yemeni etc.) tend to cut from the average noticeably. Usually known within their fam and easily visible in their 23andme results or what have you. You and me are good examples of the homogeneity found among the rest of us non rare outliers. Unrelated and only distantly tribally related yet you, me and a random Isaaq friend of mine whose sample I was fiddling with yesterday all have the same admixture level. The whole ethnic group, when you bar outliers with recent outside admixture, has a variation of
~2%.
Within the Horn among Cushites and Ethiosemites it seems to go something like this, last I recall:
Borana Oromos
>Somalis~Some Oromos
>Some Oromos~Sidamics
>Agaws~Afars
>Habeshas
>Tigres~Bejas
The least admixed of the above seem maybe ~35% Eurasian the most admixed about ~55% or so.
