Razib Khan's post at his new digs (Discover blogs) about the 17q21 inversion is worth reading for anyone interested in the complexities of discovering the history of genes in populations ("The many lives of an inverted genetic region").
The inversion in this region, common in Europeans, was described in 2005 as one of the earliest clear examples of recent positive selection determined from genotype data alone. It appeared that the apparently selected allele had diverged quite a long time ago from the wild type, leading to the hypothesis that it had sojourned in some ancient species of hominins before re-entering the human population by introgression.
But complexities followed -- for one thing, the region has repeatedly undergone inversions in other primate lineages. Now, as Razib points out, it looks like the inversion isn't all that old.
It's all very curious...
