Effect of different genotypes quality control criteria on genome-wide association studies in Nellore cattle

The aim of this study was to evaluate the effect of different criteria for the genotypes quality control on genome-wide association studies for age at first calving in Nellore cattle. Data from 1,789 females provided by a large animal breeding program were used. Animals were genotyped using the high density chip HDBovineSNP BeadChip. In the present study only samples with call rate higher than 0.90 and markers in chromosome 1 to 29 were considered. To investigate the effect of SNP quality control on genome-wide association studies three scenarios were created, varying the following criteria: genotyping rate (call rate), minor allele frequency (MAF), Hardy-Weinberg equilibrium deviation (HWE) and correlation between pairs of SNPs (r²). To construct the three scenarios, it was established threshold values for the different SNP quality control criteria, as follows: (S1) More rigorous scenario: Call rate <= 0.98; MAF <= 0.05; HWE <= 10-5 and r² >= 0.999; (S2) Less rigorous scenario: Call rate <= 0.85 and MAF <= 0.01; (S3) No SNP quality control was applied. For the scenarios S1 and S2 the quality control was executed iteratively until there were no SNPs discarded. The SNP effects were estimated by BayesC? method using phenotype adjusted for fixed effects of contemporary groups. The percentage of total genetic variance explained for each 1 Mb window was used to compare the different scenarios. The results have shown that the proportion of genetic variance explained by the same window in the different scenarios is not the same, changing the window rank. The total percentage of genetic variance explained by the top ten windows in the three scenarios were practically the same: 1.84, 1.87 and 1.85 for S1, S2 and S3, respectively. There was effect genotype quality control criteria on the genetic variance captured by each marker within the 1 Mb window.