Variance components and genetic parameters of the litter size trait, using Bayesian inference under linear and threshold animal model were estimated. The litter size of crossbred goats was studied with information regarding a period of eight consecutive years. Analyses were performed with 500,000 cycle chains. The burn-in of the 15,000 baseline values was considered and these were taken every 250 cycles to obtain a posteriori distribution with 1,940 effective samples. Statistical analyses showed that the effects of coverage month, delivery order and linear effect of weight on coverage were significant. The heritabilities were 0.03 and 0.18 for linear and threshold models respectively. The threshold model proved to be suitable, producing higher estimates regarding the estimated parameters.
bayesian inference; genetic parameters; Gibbs sampling; variance components
