Sharon Begley covers a recent paper by Joanna Mountain on Y chromosome migrations and African pastoralists:
The novel mutation arose in eastern Africa about 10,000 years ago and was carried by migration to southern Africa about 2,000 years ago not by Bantu-speakers, in whom the mutation is absent, but in speakers of what’s called the Nilotic language. These unsuspected ancestors first brought herds of animals to southern Africa before the Bantu migration.
To me, this is one of the most useful applications of genetics to prehistory: finding migrations that have been largely obscured by later movements. But it's tricky, and faces a major problem in the fact that recent selection has also generated demographic forces. Of course, if the migrations were somehow connected to the selection, that would be less of a problem...
