Non-random mate choice in humans

Are there other genetic factors that contribute to mate choice other than MHC dissimilarity? This paper published in Molecular Ecology performed a genome-wide scan of three populations of different origin (from the HapMap 3 project) - European American (CEU), Mexican (MEX), and African (YRI) - for regions of extreme similarity or dissimilarity between spouses. For the CEU and YRI populations, immunity genes exhibited excess similarity and dissimilarity; the CEU population looked to be extremely dissimilar at these loci, while the YRI sample appeared significantly similar. Additionally, two olfactory receptors were excessively similar between spouses in the MEX and YRI populations. HLA genes were not significant in this study, although they only considered 31 SNPs in comparison to over 9,000 SNPs incorporated into studies that had previously found non-random mating in relation to this locus.

Interestingly, when a functional category of genes (e.g. epidermis development) showed a strong signal in two populations, different genes were actually involved, suggesting the possibility of sexual selection occurring for the same traits in these populations, but recruiting different genes.