Goodness of Fit Comparisons among Five Bayesian Models in Genome-Wide Association of Tick Resistance in Brazilian Hereford and Braford Beef Cattle

Abstract Text: This study aimed to compare five models fitness and top effect SNPs obtained with three different Bayesian GWAS methods applied to cattle tick resistance in Braford and Hereford. After SNPs and sample’s quality control analyses, 78% of the SNPs (41,045) were selected to be used simultaneously in the GWAS analysis of 3,455 animals. Among the tested models, Bayes C (BC) was the method showing the best goodness of fit, according to the posterior mean of the log-likelihood and the Deviance Information Criterion parameters, while the worst results were obtained with Bayes B (BB3). The standardized estimated squared-marker effects and the top ten SNPs ranked across tested models also favored Bayes C method, and highlighted SNPs in BTA5, BTA11 and BTA15, especially. Results indicate that further analyses to identify specific genes or genomic regions related to cattle tick resistance should be concentrated in these chromosomes.

Keywords: genomics, marker effects, standardization.