I was reviewing some literature for a blog-post-to-come and I noticed a figure in a paper I've long been aware of which indicates to me that Afrikaners surely have a non-trivial proportion of non-European ancestry. The paper is Population differences of two coding SNPs in pigmentation-related genes SLC24A5 and SLC45A2. It's a forensics result. Basically SLC24A5 is useful for differentiating West Eurasians from Africans and East Asians, Amerindians, and Oceanians. But it is not too useful in distinguishing between West Eurasians. The "European" derived variant SNP within this locus is actually present at ~50% frequency as far south and east as India. In contrast, the "European" derived variant of SLC45A2 decreases much more rapidly outside of Europe, so it is a more plausible European-diagnostic-marker. The figure below illustrates the results from the paper:
Here's what you need to know about the samples:
Randomly selected individuals from populations sampled were Chinese from Guangzhou (South China), Uygurs from Urumqi (West China), Ghanaians from Accra, Xhosans (Africans) and European-Africans from Cape Town, and Sinhalese and Tamils from Sri Lanka. The numbers of samples analyzed were 80 for Chinese, 121 for Ghanaians, 101 for Xhosans, 101 for Europeans, 55 for Uygurs, 54 for Sinhalese, and 58 for Tamils.
The abbreviations should be self-evident now. What struck me when I thought about it is that the
ancestral allele frequency for SLC24A5 is way too high for a Northern European population
. If these are 101 whites from Cape Town, presumably they are a mix of Afrikaners (Cape Dutch variety) and English speakers. I am surprised to see any ancestral variants of SLC24A5 in a putative Northern European population. The Afrikaners are descended from Dutch, German, and Huguenot French ancestors. In the HapMap none of the 63 "Utah White" individuals (assume these are "founders") carried the ancestral allele. Of the 101 Tuscans, there was one ancestral allele copy. In the paper above they report a proportion of 0.975 for the derived European variant of SLC25A5 (probably five allele copies out of 202). Very high frequency of the derived variant, but nevertheless not as high as I would have expected. We've got a lot of data on this SNP, and the derived variant is pretty much fixed in Northern Europe, and nearly fixed in Southern Europe. A study of 158 Greeks yields two ancestral alleles. This is in the same range of the Tuscans, ~0.63%. The values for SLC45A2 alsoseem a high for a Northern Europe population, but the derived variant is not quite fixed, and the minor allele frequency is more variable across populations (the "non-European" variant is present at a less than 5% in Northern European populations, but usually more than 1%). This question will be answered definitively soon enough. There is apparently a group working on estimating admixture of non-European ancestry in Afrikaners, and the result will probably come out in the next year or so. And there has been past genealogical work which documents the intermarriage between white men at the Cape and non-European women in the early years of the colony. But, it seems likely that this admixture was substantial enough, and enough Cape Coloureds have "passed," that it is likely evident in the autosomal genome of contemporary Afrikaners. Neat to see that old data in light of all of the evidence we have about European populations can generate an unrelated inference.