THE INFLUENCE OF MATERNAL EFFECTS ON ACCURACY OF EVALUATION OF LITTERSIZE IN SWINE

The influence of maternal genetic effects on response to selection was examined by stochastic simulation. The selection process of a closed herd of 120 sows, with 24 boars entering the breeding herd each year, was simulated over 10 yr. Effects of different magnitudes of maternal heritability (0, .025, .05), genetic correlations between maternal and direct effects (0, -.5, -.9) and evaluation models (with or without m aternal effects, referred to as complete or incomplete model) on respo nse to selection, accuracy of evaluation, prediction error variance (P EV), bias, and mean squared error (MSE) were analyzed for litter size with a direct heritability of .10. Directional selection of replacemen t animals was on EBV of direct effects for first-parity litter size un der an animal model. Using a complete animal model with maternal effec ts, response in direct genetic effects increased with magnitude of mat ernal heritability (0 to .05) from 2.22 to 2.32 pigs after 10 yr, when there was no correlation between direct and maternal effects. Additio nally, a positive maternal response was achieved (with maternal herita bility > 0), although no selection was on maternal EBV. Reduction in d irect response due to negative genetic correlations between direct and maternal effects was up to 18% after 10 yr of selection. More importa nt was the negative maternal response, which was up to -1.27 pigs afte r 10 yr for a genetic correlation of -.9. Consequently, the overall ge netic merit (maternal plus direct) was reduced up to 77% compared with when maternal and direct effects were genetically independent. Estima tion of female breeding values was up to 36% more accurate than in mal es. Estimates of genetic and environmental trend, when maternal effect s were ignored in the evaluation model (incomplete model), were biased . Consideration of maternal effects in the evaluation model of first-p arity litter size was recommended.