Genome-wide association studies in admixed Latin American populations.

Abstract

Genome-Wide Association Studies (GWAS) have identified many alleles associated with human diseases and traits in the last decade. The identification of genes and causal variants for complex phenotypes is important to elucidate the genetic basis involved in the pathogenesis of diseases and to improve treatment, diagnosis and prevention. However, GWAS studies have been predominantly developed in populations of European origin. Studies in other populations are important to reveal new susceptibility loci and etiological mechanisms. In this context, the Brazilian population is of special interest due to its multiracial nature. The genotype imputation is an important step in GWAS and is the process of predicting or imputing genotypes that are not directly typed in a sample of individuals. One of its uses is to increase the power of GWAS and help combining results of studies with different genotyping platforms for meta-analysis. Nevertheless, little effort has been expended in the development of reference panels that allow robust imputation in admixed Latin American populations and few studies had touched this topic. Our goal throughout the projects was to provide more robust and effective GWAS with Latin American populations by developing an imputation reference panel for Brazilian admixed and Latin American populations and a masterscript to organize all imputation process tasks. Thus, based on data of 4.3 million SNPs from 265 admixed individuals of the EPIGEN-Brazil Initiative, we created a new imputation reference panel combining these data with 1000 Genomes Project Phase 3 data (1KGP). We then imputed SNPs from the new proposed panel on a target dataset, composed of 6487 individuals genotyped for 2.5 million SNPs, and analysed the results to compare the performance of our proposed reference panel in relation to the public panel (1KGP) available. We observed that with the EPIGEN-5M+1KGP panel were imputed 140,452 more SNPs in total and 788,873 additional SNPs with high probability values of being the correct genotypes (info score ≥ 0.8) than when using the 1KGP panel alone. Thus, the major effect of the inclusion of the EPIGEN-5M dataset in the proposition of a new imputation panel is not only to gain more SNPs but also to improve the quality of imputation. Besides that, the EPIGEN-5M+1KGP panel improves imputation quality in respect to 1KGP across a wide range of allele frequencies. We are also participating of some consortia of meta-analysis of GWAS with imputed and genotyped data from EPIGEN-Brazil Bambuí Cohort. We performed a PR interval GWAS for The CHARGE consortium and observed three important peaks at chromosomes 7, 12, and 14 in the preliminary results of regression analysis. The results will be meta-analyzed together with other GWAS.