Stated race/ethnicity is not a definitive indicator of patrilineal ancestry in males of major US populations.

Evolutionary analyses of Y-linked SNPs provide clustering of Y-haplotypes defining haplogroups, whose geographic origins have been studied at least at continental levels. Lack of recombination allows haplogroup prediction even with Y-STR haplotype data. In this research we examined variations of predicted patrilineal ancestry of individuals grouped by their self-described race/ethnicity. Y-STR haplotype data, encompassing 23 STR loci, on 936 unrelated males published by the US National Institute of Standards and Technology (NIST) were used to examine the Y-chromosomal diversity at the haplogroup level for the three self-described race/ethnicity groups, namely US Caucasians, African Americans, and US Hispanics. Of these samples, haplogroups of 814 Y-STR haplotypes were predicted with a Bayesian approach, from which haplogoup diversity was estimated for the three groups. The phenetic tree of the observed Y-STR haplotypes was drawn as a Median Joining Network (MJN) with the program Network, on which their haplogroup prediction and stated race/ethnicity information was superimposed to graphically demonstrate the differences of patrilineal ancestry among the three groups. These predictions were used to estimate the proportions of male ancestry from European, African, and Native American gene pools. Predominantly most (90.85%) US Caucasians had predicted male European ancestry, while the Hispanics and African-Americans had much lower levels of male European ancestry; 62.95% and 31.14%, respectively. Contributions of male African ancestry in Caucasians, Hispanics and African-Americans were approximately 4.27%, 15.23% and 67.47% , respectively. Lastly, all three groups exhibited variable amounts of male Native American ancestry; namely, 0.31%, 0.69%, and 14.72%, respectively for Caucasians, African-Americans, and Hispanics. At the haplogroup level, the coefficient of haplogroup diversity (GST) was 18.15%, suggesting that the three groups differed substantially with respect their male ancestral haplogroup lineages. In contrast, these three groups had 812 distinct 23-locus haplotypes; none of which were shared across the three groups. Only two haplotypes were observed twice each, one in African-Americans, and one in Hispanics. The 328 Caucasians had all distinct haplotypes. These resulted in a coefficient of haplotype diversity (GST) of 0.26% at the haplotype level. Together these results suggest that while there is virtually no haplotype sharing across these three groups, each of them exhibited considerable amount of haplogroup sharing, and hence prediction of male lineages is not definitive from the stated race/ethnicity of individuals with Y-STR haplotypes defined by the 23 loci of this study.