Application of the F ST statistics to explore pharmacogenomic diversity in the Brazilian population

BACKGROUND: New drug applications submitted to regulatory agencies in developing countries rarely include data from local clinical trials. We used the F(ST) statistics to explore the pharmacogenomic diversity of the Brazilian population and its potential implications in drug regulatory assessment and decisions. METHODS: The F(ST) analyses were based on data for 44 polymorphisms in 12 pharmacogenes among 1034 healthy Brazilians, recruited in four different geographical regions and self-identified as branco (white) pardo (brown) or preto (black). Each region/color group comprised 83-89 individuals. The Utah residents of northern and western European ancestry and Yoruba people from Nigeria, Africa, cohorts of the HapMap project were used as proxies of the European and sub-Saharan African ancestral roots of Brazilians, respectively. RESULTS: Allele-specific F(ST) values for the overall Brazilian cohort revealed low genetic divergence between white and brown (F(ST) = 0.005 ± 0.006, mean ± standard deviation), white and black (0.013 ± 0.017) and brown and black (0.004 ± 0.005) individuals. However, the distribution of F(ST) values for white vs brown (p < 0.0001, analysis of variance) and white vs black (p < 0.0001) differed significantly across the geographical regions. Considerably larger pharmacogenomic divergence was observed between black Brazilians and Yoruba people from Nigeria, Africa (F(ST) = 0.028 ± 0.035) compared to white Brazilians vs Utah residents of northern and western European ancestry (0.007 ± 0.010). CONCLUSION: The present F(ST) analyses highlight the challenge faced by Brazilian regulatory agencies when assessing the relevance to Brazilians of pharmacogenomic data derived from foreign populations, with distinct biogeographical ancestries. This challenge is compounded by the heterogeneity of the Brazilian population with respect to the frequency distribution of pharmacogenomic polymorphisms across color categories and geographical regions.