Bayesian inference of variance components for litter size in Rasa Aragonesa sheep

Variance components were estimated for litter size in Rasa Aragonesa sheep, a meat breed from northern Spain, to determine whether selective breeding for litter size is a reasonable strategy to improve reproductive performance. We assumed an animal mixed effect threshold model with a binary response variable. Marginal estimates of the genetic parameters were obtained in the underlying scale using Bayesian inference, implemented via the Gibbs sampling procedure and a data augmentation approach. Posterior marginal means of heritability and repeatability were .077 and .141, respectively. Moreover, the 95% highest marginal posterior density region of heritability ranged from .051 to .101. Therefore, we conclude that litter size is a trait that could be selected for in breeding programs. The effect of the loss of pedigree information, a common feature of sheep production, on the estimation of the genetic parameters was also studied using simulation. The results indicate that the lack of pedigree information has little effect on our estimates of heritability.