RESUMO
Narrowing of genetic diversity and the quantitative nature of most agronomic traits is a challenge for rice breeding. Genome-wide association studies have a great potential to identify important variation in loci underlying quantitative and complex traits; however, before performing the analysis, it is important to assess parameters of the genotypic data and population under study, to improve the accuracy of the genotype-phenotype associations. The aim of this study was to access the genetic diversity, linkage disequilibrium, and population structure of a working panel of Brazilian and several introduced rice accessions, which are currently being phenotyped for a vast number of traits to undergo association mapping. Ninety-four accessions were genotyped with 7098 SNPs, and after filtering for higher call rates and removing rare variants, 93 accessions and 4973 high-quality SNPs remained for subsequent analyses and association studies. The overall mean of the polymorphic information content, heterozygosity, and gene diversity of the SNPs was comparable to other rice panels. The r2 measure of linkage disequilibrium decayed to 0.25 in approximately 150 kb, a slow decay, explained by the autogamous nature of rice and the small size of the panel. Regarding population structure, eight groups were formed according to Bayesian clustering. Principle components and neighbor-joining analyses were able to distinguish part of the groups formed, mainly regarding the sub-species indica and japonica. Our results demonstrate that the population and SNPs are of high quality for association mapping.