WORLDWIDE GENETIC DIVERSITY AT THE HLA-DQA1 LOCUS

Genotype, gene, or phenotype frequency data, obtained by PCR analysis with sequence specific oligonucleotide probes at the HLA-DQA1 locus, i n 176 population samples, each consisting of 30 or more individuals, f rom all around the world were analyzed. The sampled populations were a ssigned to the following nine groups: African, Afro-American, American Native, American Mestizo, Asian, American Caucasian, European Caucasi an, Caucasian from other regions, and Pacific Islanders. Observed geno type proportions agreed with Hardy-Weinberg expectations (HWE) in 90 o f the 102 populations for which genotype data are available. The 12 di scordant population samples were of substantial mixed origin. For thes e discordant populations, 41 of the 252 (population by genotype) data points showed significant departures of the observed frequencies in co mparison to their HWE predictions. The deviations, by and large (38 of the 41), were in the direction of HWE over-estimating the actual geno type frequencies. Total heterozygosity (H-T) at a worldwide level was 79%, and varied from 56% in American Natives to >80% in Caucasians. Of the total diversity, 94.4% was due to intrapopulation and 5.6% to int er-population variation (F-ST), partitioned into 3.0% inter-population within groups and 2.6% inter-group variation. Fs, for the locus was h igh among Africans and American Natives (>9%) and low in Afro-American s and American Caucasians (<0.5%). The range for allele specific F-ST was 0.2% to 5.9%, and the lowest value did not correspond to the same allele in all nine groups. A variance component analysis of allele fre quencies showed no relationship of the ratio of between/within group v ariation with the world average frequency of the respective alleles. T he first three principal components explained 36.2%, 27.5%, and 22.9% of the total allelic diversity, respectively. For Caucasian and Afro-A merican samples, the first two PCs formed clusters by groups. In contr ast, the American Native, Asian, and Pacific Islander groups showed a greater inter-population diversity, while the corresponding principal component scores of the American Mestizo samples were between American Natives and Caucasians. In aggregate, the analyses indicate that gene tic drift in contrast to natural selection, more readily explains the pattern of worldwide diversity at the HLA-DQA1 locus. (C) Wiley-Liss, Inc.