Genetic analysis of birth weight uniformity in pigs: comparison of methods

Abstract Text: Uniformity is an important goal in animal breeding, but selection for uniformity is limited to traits with standard deviation or variance. The objective was the evaluation of Double Hierarchical GLM (DHGLM) as a tool to study uniformity of traits with a single observation and its application to birth weight (BW) in pigs, treated as sow’s observation. DHGLM studies differences between animals by analyzing the residual variance of trait and estimating its variance components in ASReml. Comparison with traditional method using log-transformed variance of BW was possible, because methods are on the same scale and contain same random effects. Both methods produced highly comparable estimates of variance components, EBVs and accuracies of EBVs, and showed similar response to selection in BW variation. DHGLM has potential to be used to study the uniformity of traits with a single observation.

Keywords: Birth weight variation, Double hierarchical GLM, Heterogeneity of variance, Pigs