Modeling breed additive and non-additive genetic effects using a Angus x Nellore crossbred population

The objectives of this study were to estimate fixed genetic (both additive and non-additive) effects and test different non-additive models. These effects included direct and maternal joint breed additive effects, direct and maternal complementarity (or profit heterosis), direct and maternal heterosis (dominance) and direct and maternal epistatic loss non-additive effects. A large crossbred Angus x Nellore population with approximately 300,000 records for weaning gain (WG) and 150,000 records for post-weaning gain (PG) was used. Phenotypic scores for weaning (WC) and post-weaning (PC) conformation, weaning (WP) and post-weaning (PP) precocity, weaning (WM) and post-weaning (PM) muscling and scrotal circumference (SC) were also used. All models included the fixed contemporary group effect and random animal, maternal genetic and permanent environment effects. Each model was tested against all other models for all nine traits using the likelihood ratio test. The complete model, including all additive and non-additive effects and the model without complementarity, were seen to be the best options to analyze this crossbred population. However, least squares may not be the best methodology due to possible collinearity among estimates that may inflate the variance. In the complete model most effects were statistically significant (P<0.01) for weaning traits, except for direct and maternal breed additive effects and direct complementarity effect for WM. For post-weaning traits, the direct complementarity effect for phenotypic scores (PC, PP and PM) and the maternal heterosis effect for PG, PC, PM, were not statistically significant. For scrotal circumference the maternal complementarity, direct and maternal heterosis and maternal epistatic loss effects were not statistically significant. All other effects were statistically significant (P<0.01). For the model without complementarity, the direct breed additive effect for PG and PP was not statistically significant, including the maternal breed additive effect for WC, WD, PP and PM and maternal heterosis, direct and maternal epistatic loss effects for SC. Considering all models, significant direct breed additive effect was mostly positive for weaning traits and was negative for all post-weaning and SC. The significant maternal breed additive and maternal complementarity effects for SC were all negative. The significant direct and maternal epistatic loss effects were negative for all traits and for all models, except for PG in the complete model and for weaning traits in the model including only breed additive and epistatic loss effects. We conclude that the fixed genetic effects are mostly significant, thus it is important to include them in the model when evaluating crossbreed animals and the models including breed additive effects, heterosis and epistatic loss with or without complementarity were the more appropriate.