Inferência bayesiana na predição de valores genéticos do peso aos 365 dias de bovinos de corte
Título alternativoBayesian inference to predict genetic values for 365 days weight in Simmental, Nellore and Canchim cattle breeds
Data de publicação2009-06-01
Direito de acesso
MetadadosExibir registro completo
Data from purebred Simmental, Nellore and Canchim cattle breeds obtained from the respective Brazilian Associations of Breeders were used to estimate variance components and to predict genetic values for 365 days weight. The results obtained by Bayesian inference were compared to those from Restricted Maximum Likelihood (REML) and Best Linear Unbiased Prediction (BLUP), which are the most commonly used methods of estimation and prediction in animal breeding. The two methods presented similar point estimates but the study of the marginal posterior distributions in the Bayesian approach yields more detailed information about the parameters and other unknowns in the model.