|Location of the samples|
There is a new paper dealing with Sudanese genetics, which is of some interest, in my opinion:
Hima M.A. Babiker et al., Genetic variation and population structure among Sudanese populations as indicated by the 15 Identifiler STR loci. BMC Investigative Genetics 2011. Open access.
Most interestingly they spot a cline and divide (both) between Egypt and Tropical Africa and differential genetics for Somalis, one of three outgroups (the other two being Egyptians and Ugandans from Karamoja region (Nilotic ethnicities).
|Fig. 5 (blue: my annotations)|
The structure shows (as much as a mere three components allow) a duality of some clinal value (i.e. not absolute but relative) with a divide cutting across Sudan: not just between North and South according to the latest political split but also placing Darfur, Kordofan and the Nuba along with Tropical Africa as well.
Instead Central, North and Eastern Sudanese look much like Egyptians. Somalis are clearly different however but you'd never know based only on K=2. That's why it is important to explore these analysis to some depth, greater than K=3 in any case, a very shallow depth for such a diverse region.
Notice that PC2 and PC3 are of similar values. PC1 however is more than double in importance and marks a cline between Egypt (and the Sudanese Copts) and the Nuba. PC2 and PC3 only show distinctions between Copts and Egyptians and Copts and Somalis respectively.
I wonder if this last is caused because of random peculiarities of the 15 ancestry informative markers used in this study, which seem a bit too few not to cause random distortions, specially in such a poorly understood region as is East Africa.