In 2008 a paper came out which reported that Europeans had more genetic load than Africans, Proportionally more deleterious genetic variation in European than in African populations. At the time I recall Greg Cochran was somewhat skeptical on grounds of biomedicine, and some rather unrealistic demographic assumptions (an realistically long bottleneck). The basic finding was simple, because of the “Out of Africa” event Europeans (and presumably all non-Africans) would exhibit a higher load of deleterious alleles because of the reduced power of selection in relation to drift. Over the past seven years that simple result has come under critique, and the first author of the 2008 paper now has a review which resolves the conflicting results, The distribution of deleterious genetic variation in human populations, out (the link is to the preprint, which has been around for a while). The short of it seems to be that the distribution of frequencies of deleterious alleles may differ across populations as a function of demographic history, with the bottleneck and rapid population growth resulted in an excess of rare alleles in non-Africans, but the large population producing more efficacy of selection. The theory itself in the paper is less interesting to me than the conclusion. Here he states:
Future work should include examining empirical patterns of deleterious mutations in other human populations that have differing populations histories, such as different amounts of recent population growth. Studies with large samples of individuals will be particularly helpful as they will be informative regarding how deleterious mutations have behaved during recent times….
Genomics is powerful. For the sort of subtle evolutionary patterns which researchers are trying to sniff out it strikes me that good quality whole genomes in larger numbers across more populations are probably necessary before we can make robust generalizations about humans, let alone other species. Cautious is definitely important because the first wave of SNP-chip results seem to have produced a set of results which were interpreted in light of theory, without understanding that the empirical results were only a sliver of reality constrained by the methods at hand.