I'm intentionally giving this an unexciting title because I'm disappointed. There's a great-sounding title in this month's Trends in Genetics: Genes of domestic mammals augmented by backcrossing , which unfortunately lacks a punch.
Sequencing of mitochondrial DNA (which is only inherited from mothers) from most domesticated mammals- cattle, sheep, goats, pigs, dogs - indicates that each species was probably adopted from the wild only a few times. This is in agreement (Wikipedia) with archeological data. The note of the current paper is that other DNA regions (in this case parts of the MHC locus) are far more diverse and require a bigger founding population. The proposed modification of the existing theory is that genetic diversity was augmented by crossing in wild animals (Males? They would not contribute to mtDNA) into the domesticated group.
I just don't think this should suprise anyone, not even the ones who published the mtDNA surveys. Of course there would be a transition period in domestications in which genetic flow was a bit more, well, fluid. Especially if the animals tend to end up as dinner, a backcross may have been the way to keep herd sizes reasonable.
What would get me excited would have been to map out regional MHC diversity- to take an extreme example, did Chinese pigs get backcrossed with Chinese boars, and British with British? Detection of "wild" MHC alleles in local domestic populations would have been a much more direct demonstration of the principle. DNA should not be that hard to come by..