If you follow Y genealogy you know that the distribution of R1ba2 exhibits a peculiar pattern. R1b is the most common haplgroup in Western Eurasia, and shares a deep common ancestry with R1a. It seems to have risen to high frequencies in Europe only during the Bronze Age, though has been found in earlier periods. But within Africa R1b is found in very high concentrations around Lake Chad. This particular R1b lineage seems to have diverged from other Eurasian branches in the latter portion of the Pleistocene, so one possible consideration is that this was an instance of Eurasian backflow during the Ice Age.
One reason I have been somewhat skeptical of this model is that the Sahara desert was much more extensive and arid during much of the Pleistocene than today. And during this period humans had less cultural technology to endure the rigors of the deep desert. Or, if they did, their population densities were likely much lower, which probably served as an impediment to gene flow.
A new paper in The American Journal of Human Genetics sheds light on what might have been going on here. Chad Genetic Diversity Reveals an African History Marked by Multiple Holocene Eurasian Migrations. The major findings are straightforward. First, much greater sampling of populations, and a better depth/density of marker coverage, allowed the researchers to detect low levels, on the order of ~1%, Eurasian admixture in some Central African groups. This admixture seems to date to the Holocene, ~5,000 to ~7,000 years before the present (they used LD based methods on the autosome). Interestingly, the R1b lineage common in Central Africa also seems to coalesce during this time. Finally, the admixture seems to be closest to Sardinians among extant populations.
The Sardinian affinity of much of African Eurasian admixture may seem peculiar, but it makes more sense when one considers that Sardianians are probably the best modern proxies for the earliest Neolithic farmers from the Eastern Mediterranean. Modern Middle Eastern populations are very different from those which flourished in the prehistory between the rise of agriculture and complex civilizations because of admixture within Middle Eastern groups. The initial push into Africa by the agriculturalists dates to a period before we have a good understanding of the ethnographic balance.
Very high frequencies of R1b in modern Central Africa groups may indicate drift. But another possibility is that the migration was male-mediated. This seems to have been the case in much of Eurasia, so it would not be surprising in this context. The status of these males was such that despite their diminishing genetic impact on overall ancestry, their Y chromosomes, and possibly their language, with varied forms of Afro-Asiatic, persisting down to the present.
Finally, here’s the last paragraph of the discussion:
Our study has shown that human genetic diversity in Africa is still incompletely understood and that ancient admixture adds to its complexity. This work highlights the importance of exploring underrepresented populations, such as those from Chad, in genetic studies to improve our understanding of the demographic processes that shaped genetic variation in Africa and globally.