Diversity at eight polymorphic Alu insertion loci in Chinese populations shows evidence for European admixture in an ethnic minority population from Northwest China

Human Biology, Aug 2002 by Xiao, Feng-Xia, Yang, Jun-Fang, Cassiman, Jean-Jacques, Decorte, Ronny

As the ancestral state of these Alu insertion polymorphisms is known with certainty to be the absence of an Alu element, we rooted the tree with a hypothetical ancestral population in which the frequency of the Alu element at each locus was set to zero. Inclusion of this ancestral population placed the root within the cluster of African populations as indicated by the arrow in Figure 2. This provides additional evidence for the support of the "out-of-Africa" hypothesis of modern humans (Cann et al. 1987; Bowcock et al. 1994).

The population relationship was also calculated using a principal component analysis (Cavalli-Sforza et al. 1994) of the allele frequencies at the eight loci (Figure 3). The first two principal components, which account for 83% of the variation in allele frequencies, were used to plot the positions of the populations. All the African populations are closest to the Ancestral population in the first PC, which would be consistent with the African origin of modern humans. The second PC placed the population relationships similar to those observed in the NJ tree. However, geographic clusters of the East Asian populations were not as evident as in the NJ tree, in which the India-Christian group was close to the East Asian populations, whereas the Filipinos were close to the two West Asian populations (India-Hindu and Tamil), as were the Uyghurs. The Pakistani and United Arab Emirates (UAE) populations were found to lie between the Europeans and other West Asians, showing more concordance with geography than the NJ tree. The Uyghurs approximately fell between the East Asians and the West Asians, which supports the admixture scenario for this population. When the hypothetical ancestral population, with the allele frequency at each locus set to zero, was removed, the same pattern of population distributions was observed.

Gene Flow between Populations. To estimate the relative amount of gene flow experienced by each population studied, we plotted the heterozygosity of each population against the distance of the population from the centroid (the overall mean allele frequencies of the populations) (Harpending and Ward 1982). This model assumes a linear relationship between the heterozygosity of a population and the distance of the population from the centroid. In a population that has received more gene flow than average, the heterozygosity will be higher than predicted by the model; in a more isolated population, the heterozygosity will be lower than predicted. In Figure 4A, the two minority populations, especially the Uyghurs, showed a greater heterozygosity than predicted, whereas the two Han populations had a lower heterozygosity than predicted. This implies that the two minority populations have received more gene flow than average, while the two Han groups are more isolated.

We also performed this centroid analysis by pooling the data of the present study with the 29 worldwide populations (Stoneking et al. 1997) included in the phylogenetic analysis (Figure 4B). All six African and two West Asian populations (Tamil and India-Christians) showed greatly higher heterozygosities than predicted, while all four Chinese populations exhibited lower values than predicted. The greater African heterozygosities have been interpreted as evidence supporting the "out-of-Africa" hypotheses and a larger effective population size in Africa (Stoneking et al. 1997). Inclusion of the four Chinese populations in the centroid analysis showed again a greater heterozygosity in African populations, which led to the further support for the African origin of modern humans in East Asia.