Tuesday, January 23, 2007

Where are the freaks?   posted by Razib @ 1/23/2007 10:53:00 PM

So I've been mulling over the recent publication in Annals of Human Genetics of a review of the recent skin color genomic work. The conclusion is pretty predictable given the recent findings:

a) Dark skin is the modern human ancestral trait
b) Light skin is derived
c) The derivations are independent

There is lots of stuff to comment on, but I'll limit myself to a weird thought I've had for a while. The authors point out that East and West Eurasians (e.g., Western Europeans and Chinese) are light, in general, because of different mutations on different loci. In other words, the genetic architecture is pretty dissimilar. Even in the one case where the same locus (or genomic region) was subject to selection the haplotype differed. One would expect that there would be overlap in some of these genes being selected for since they are implicated in the same phenotype, though the allelic solution was distinct. Nevertheless, my interest is in the loci which do not overlap (most). Consider SLC24A5. It explains around 30% of the intergroup variance between Europeans and Africans, but none of the variance between East Asians and Africans, because East Asians and Africans share the ancestral allele. In contrast, MC1R is hyperpolymorphic in Europeans, constrained to the ancestral state in Africans, and being positively selected in East Asians toward fixation. And so on. Now...imagine, you have loci:

1, 2, 3, 4, 5, 6, 7, 8

...implicated in the loss of melanin production in human skin. Europeans are derived on:

1, 2, 3, 4, 5 (so ancestral on 6, 7, 8)

East Asians on

5, 6, 7, 8 (so ancestral on 1, 2, 3, 4)

Assuming that the loci are fixed, if you crossed a bunch of Asians with a bunch of Europeans (here's looking at you Hawaii!), after a few generations you could have someone who is derived on:
1, 2, 3, 4, 5, 6, 7, 8 homozygously

Greg points out that these selected genes seem to be relatively recent (agricultural?), so their shallowness means they aren't embedded in coadapted complexes which are likely to birth monsters. In fact, we know from pedigree studies that between Europeans and Africans skin color is inherited pretty much in an independent and additive fashion with 4-5 loci accounting for 90% of the between racial variance. So I am wondering if any intrepid readers want to engage in skin reflectance tests of variously racially mixed happas in Hawaii?