Estimated breeding values for number born alive in pigs - do they really work?

Estimated breeding values (EBVs) have been used in pig genetic improvement programs for several years. Reproductive traits, such as number born alive (NBA), have only low to moderate heritability (et al., 2005). Nevertheless, using EBVs, selection for NBA in maternal lines at the seedstock level has resulted in genetic improvement at the purebred level, which should flow to the commercial level in the next generation. Progeny receive half their genes from their dam so EBVs for NBA of purebred dams should predict half the average differences in performance for litter size of their daughters. Rather than using EBV, some producers instead rely on the phenotypic performance of the dam as it predictor of genetic merit for litter size. The aim of this paper was to investigate whether EBVs for NBA of dams at the purebred level predict performance of crossbred daughters at the commercial level. In addition, the ability of EBVs and phenotypic performance of dams to predict NBA was compared.