Mexican-American Ancestry-Informative Markers

Human Genetics
DOI: 10.1007/s00439-003-1058-6

Mexican American ancestry-informative markers: examination of population structure and marker characteristics in European Americans, Mexican Americans, Amerindians and Asians

Heather E. Collins-Schramm et al.

Abstract Markers with large differences in allele frequencies between ethnicities provide ancestry information that can be applied to genetic studies. We identified over 100 biallelic ancestry informative markers (AIMs) with large allele frequency differences between European Americans (EA) and Pima Amerindians from laboratory and database screens. For 35 of these markers, Mayan, Yavapai and Quechuan Amerindians were genotyped and compared with EA and Pima allele frequencies. Markers with large allele frequency differences between EA and one Amerindian tribe showed only small differences between the Amerindian tribes. Examination of structure in individuals demonstrated a clear separation of subjects of European from those of Amerindian ancestry, and similarity between individuals from disparate Amerindian populations. The AIMs demonstrated the variation in ancestral composition of individual Mexican Americans, providing evidence of applicability in admixture mapping and in controlling for structure in association tests. In addition, a high percentage of single-nucleotide polymorphisms (SNPs) selected on the basis of large frequency differences between EA and Asian populations had large allele frequency differences between EA and Amerindians, suggesting an efficient method for greatly expanding AIMs for use in admixture mapping/structure analysis in Mexican Americans. Together, these data provide additional support for the practical application of admixture mapping in the Mexican American population.

...

As a first step in examining the relationships between EA and AI, a panel of 35 indels with large differences between EA and Pima were chosen. These markers were previously identified in our screen for EA-AI AIMs, and have an average EA-Pima f of 0.36 (range, 0.18–0.59). The markers were genotyped in EA, MA, and four AI populations: Mayan, Pima, Quechuan, and Yavapai. Each AI population was compared with EA at the level of the individual using the program STRUCTURE (Fig. 1). Each marker was in Hardy-Weinberg equilibrium in each population. For each comparison the STRUCTURE analysis correctly separated EA from AI individuals without prior knowledge of population membership. In each case the overall population separation is clear, with the majority of EA identified as being 80% or more derived from population 1 and the majority of AI being 80% or more derived from population 2. However, the figure also illustrates that each population has apparent outliers. Examination of the dataset by box test (see Materials and methods) confirmed this finding. There are two outliers for Mayan, three for Pima, and six each for Quechuan and Yavapai. There was slight variation in EA outliers identified by each comparison. When compared with Mayan, there were nine EA outliers, when compared with Yavapai there were eight, and when compared with Pima or Quechuan there were seven. However, six of these individuals were outliers in all comparisons. The AI outliers and the six consistent EA outliers may represent unreported recent admixture with these or other (e.g., African or Asian) ethnic groups. These individuals were removed from all subsequent analyses.

...

The MA individuals examined in this study resided in Northern California and defined themselves as Mexican American. Thus, this population is slightly different than other Hispanic populations described to date, which include individuals from Arizona and from Texas (Chakraborty and Weiss 1986; Long et al. 1991). The average admixture estimates for these MA range from 22% AI to 50% AI. MA individuals in the present study had AI contributions varying from 10 to 80% AI, suggesting that there is likely to be at least as much variability in admixture ratios within a MA sub-population as between them. This stresses the importance of methods to measure and account for admixture ratios of each individual in admixture mapping.

Posted by Dienekes at November 28, 2003 12:30 AM | PermaLink