BAYESIAN INFERENCE IN THE GENETIC EVALUATION OF THE BRAZILIAN NORTHEAST TABAPUÃ CATTLE

ABSTRACT: The objective was to compare models by Bayes factor (BF) to estimate the (co)variance components, genetic parameters and genetic values for the adjusted weights at 205 (W205), 365 (W365) and 550 (W550) days of age in Tabapuã animals reared on pasture in the Brazilian Northeast. Information about 26,838 cattle born in the period from 1975 to 2007 originally from the weight-development-control system of the Brazilian Association of Zebu Farmers (Associação Brasileira de Criadores de Zebu, ABCZ) was used. The (co)variance components and genetic parameters were estimated by the Bayesian approach, in single- and two-trait analysis, using the GIBBS3F90 software. The full model provided the best fit to estimate the breeding value for W205 and W365 characteristics parameters. For trait W550, the model that provided the best fit was the one that did not include the effect of permanent environment. Genetic trends of direct effect were significant for the W205, W365 and W550 features, corresponding to increments of 2.67 kg, 7.35 kg and 8.32 kg, respectively, during the 32 years evaluated. Phenotypic trends were significant only for W205 and W365 characteristics corresponding to increments of 26.82 kg and 22.42 kg, respectively. The genetic correlations between weight at 205, 365 and 550 days of age were of high magnitude (P205P365 = 0.90, = 0.90 and P205P550 P365P550 = 0.98); you can obtain genetic gains via correlated response to the weights at 365 and 550 days old when selecting for W205.