Why race as a biological construct matters

Razib Khan in Gene Expression:

Gb-2009-10-12-r141-1My own inclination has been to not get bogged down in the latest race and IQ controversy because I don’t have that much time, and the core readership here is probably not going to get any new information from me, since this is not an area of hot novel research. But that doesn’t mean the rest of the world isn’t talking, and I think perhaps it might be useful for people if I stepped a bit into this discussion between Andrew Sullivan and Ta-Nehisi Coates specifically. My primary concern is that here we have two literary intellectuals arguing about a complex topic which spans the humanities andthe sciences. Ta-Nehisi, as one who studies history, feels confident that he can dismiss the utility of racial population structure categorization because as he says, “no coherent, fixed definition of race actually exists.” I am actually more of a history guy than a math guy, not because I love history more than math, but because I am not very good at math. And I’ve even read books such as The Rise and Fall of the Caucasian Race and The History of White People (as well as biographies of older racial theorists, such as Madison Grant). So I am not entirely ignorant of Ta-Nehisi’s bailiwick, but, I think it would be prudent for the hoarders of old texts to become a touch more familiar with the crisp formalities of the natural sciences.

In his posts on this topic Ta-Nehisi repeatedly points to the real diversity in physical type and ancestry among African Americans, despite acknowledging implicitly the shared preponderant history. But today with genomic methods we have a rather better idea of the distribution of ancestry among African Americans. The above plot is from Characterizing the admixed African ancestry of African Americans, a 2009 paper with 94 Africans of diverse geographic origins, 136 African Americans, and 38 European Americans. They looked at 450,000 genetic variants (SNPs) per person (there are somewhat more than 10 million SNPs in the human genome). Obviously individuals and populations exhibit genetic relationships to each other contingent upon the patterns of the variation of base pairs (A, C, G, and T) across the genomes of individuals, but there’s no reasonable way to comprehend this “by eye” when you’re talking about hundreds of thousands of markers. The authors used two simple methods to infer clustering within the data set.

More here.