Abortion & blood groups

On page 380 of Population Genetics and Microevolutionary Theory I stumbled onto some interesting data from the 1970s. The paper was Admixture Studies and the Detection of Selection , and its aim was to gage the extent of white ancestry within black populations. They generated a measure, M, for the extent of ancestry across a range of loci like so:

change in allele frequency between African Americans and West Africans
____________________________________________________________

change in allele frequency between Europeans and West Africans

In short M relates the ratio of the genetic distance between African Americans and West Africans and African Americans and Europeans. Since one presumes Africans Americans are mostly West African, with a minor component of European ancestry, one would expect that the values would be below 1, approaching, but not equal to, 0. Some results are presented below the fold.

The first six results are pretty unsurprising. Values in the range of 0.10 and 0.20 indicate European ancestry in the range of 10-20%. The relatively low values for the population from Georgia is also understandable. Modern genomic techniques usually imply that in the national black American population around 20% of the ancestry is white, but the values are closer to 10% or even below in some groups from Georgia. But what about the other results?

Hp¹, GP6PD A^- and β-Hb S are loci where there are malaria adaptations. The results from these genes suggests that around half of their ancestry is European. What's going on here? Selection. Since the 150-300 years (6-12 generations) that the ancestors of these black Americans left Africa the adaptations which confer defense against malaria have been relatively useless, and even deleterious. So alleles from Europeans, or ancestral alleles within the African population, increased in frequency subject to positive selection. This shows the power of selection to reshape extant genetic variation on a historical time scale.

But what about A? This is the A blood group. The genetics of the ABO system is pretty simple. There's one gene, and three flavors of alleles, ones which produce the A antigen, ones which produce the B antigen, and those which produce no antigen, O. If you have an A allele, you produce A, and if you have a B, you have that variety of antigen. The only way you can have blood group AB is to inherit A and B, one from each parent. Myself, I'm A. That means I have at least one copy of A, but I could have an O or an A on the other copy. I know that my mother is an A, my father a B. My father can't be a BB, because I don't have a B, he must be BO, and he passed his O to me. My mother could be AA or AO, but I know she's AO, because I have a brother who is a B. He has to be BO, while I'm an AO. In other words, A and B are codominant, and O is recessive.

OK, enough simple high school genetics. What does an M value of -0.037 mean? It is about zero when statistical error is taken into account, and would imply basically no admixture of European genes into the African population. We know this isn't true, from both these data and a host of other studies. Well, it turns out there is some data that mothers who manifest the O blood group (OO) who carry A blood group fetuses (AO, they received the A from the father) might be subject to greater rates of spontaneous abortion because of maternal-fetal incompatibilities. Similar results seem to apply to matings of women with blood type B and men with blood type A (of both parents are heterozygous then of course in 1/4 of the cases they will produce O blood group offspring). It turns out that West African populations have higher frequencies of O and B than European populations. The historical pattern (confirmed by studies of uniparental lineages transmitted from only one sex) was for European men to produce offspring with African women. The spontaneous abortion of fetuses who carried the A allele is then explicable. What you have here is viability selection that prevents the penetration of one population by another via the dynamic of gene flow! This makes me wonder about other viability barriers there might be between populations. After all, geneticists haven't found many fitness implications in hybrid matings, but this could be due to the fact that most of the incompatibilities are purged during the stage of gestation (remember that an enormous proportion of fertilizations, anywhere from 1/4 to 3/4 depending on what research you read, end in spontaneous abortions).