The title is provocative, but I just want to draw attention to Razib's comment on the new paper.
In addition I want to clarify I intend to go further on why there are so many discrepancies in the Masai vs Yoruba North African runs soon. This may be confusing to many, but I believe results are telling us something.
One reason is due to different pole sizes. I have very many Yoruba in my database (some 150) and in some runs I used only 5 in others 24 of them, creating quite different results versus the full set in supervised mode. ADMIXTURE stretches small poles a lot more than large ones, sometimes making them about something different from the pole individual's population.
This indicates there may have been a Northwest African aboriginal population quite distinct from modern West and East Africans, but related to them. This component even sometimes appears as if more distant from Eurasians than West Africans themselves...
Also high absorption levels in a Mediterranean climate environment IMO indicates some incipient level of Neolithic adaptation in this people before partial replacement by Mesopotamian/Anatolian/Levant and Northeast African Neolithics.