Objective: To compare the effectiveness of self-identified ethnic/color and marker-based biogeographical ancestry classifications in genotyping the CYP3A5*3 polymorphism in the Brazilian population.
Methods: Individual DNA from 308 healthy Brazilians, self-identified as white, intermediate and black was genotyped for the CYP3A5*3 polymorphism and for a set of insertion-deletion polymorphisms, validated as ancestry informative markers (AIMs). The Structure software was used to analyze the AIMs data and to obtain estimates of the African component of ancestry (ACA). Nonlinear logistic regression modeling was developed to describe the association between the CYP3A5*3 polymorphism and the individual ACA values.
Results: The CYP3A5*3 allele and genotype distribution differed significantly across the self-reported 'color' groups (p < 0.0001, Fisher exact test), with a trend for decreasing frequency of both the CYP3A5*3 allele and the *3/*3 genotype from white to intermediate to black individuals (p < 0.0001, chi(2) test for trend in proportions). When the population sample was proportioned in quartiles according to the individual ACA values, the frequency of the CYP3A5*3 allele and the *3/*3 genotype declined progressively from the lowest (<0.25 ACA) to the highest (>0.75 ACA) quartile. Nonlinear logistic regression showed that the odds of having the CYP3A5*3 allele decreases monotonically (p < 0.0001, Wald statistics) with the increase of the ACA, throughout the ACA range (0.15-0.93) observed in the overall population sample.
Conclusion: Interethnic admixture is a source of cryptic population structure that may lead to spurious genotype-phenotype associations in pharmacogenetic/-genomic studies. Logistic regression modeling of CYP3A5*3 polymorphism shows that admixture must be dealt with as a continuous variable, rather than proportioned in arbitrary subcategories for the convenience of data quantification and analysis.