Efeito materno sobre a curva de crescimento de ovinos Santa Inês por meio de modelos de regressão aleatória

Abstract

Data set of 17.767 weight records of 4.210 Santa Inês lambs were used aiming to evaluate the importance of the inclusion of the maternal effect in the model to estimate components of (Co) variance and resulting genetic parameters for the growth curve through random regression models. The fixed and random regressions were fitted using Legendre Polynomials of order three, being fit four models that differed in relation to the inclusion of the additive genetic and permanent environmental maternal effects. Considerable increase was observed in Log L and decrease in the criteria AIC and BIC when the maternal effect was included (genetic or permanent environmental), evidencing its importance. The maternal genetic effect explained larger proportion of the phenotypic variance than the maternal permanent environmental along the growth curve. The direct additive genetic variance was inflated by maternal effect, when this last one was not considered in the analysis model, reflecting the same behavior in the heritabilities. The maternal permanent environmental effect contributed to maternal variance, as well as, it inflated maternal genetic variance, when it was not considered in the model. Similar behavior was verified with maternal heritability. The correlation estimated for the four models hardly differed in function of maternal effect. The maternal effect should be considered in the genetic studies of the growth curve of Santa Inês sheep.