(Co)variance Components for Birth and Weaning Weights of Shorthorn Beef Cattle in Australia and the United States

Abstract

(Co)variance components and genetic parameters for birth (BW) and weaning weights (WW) of Shorthorn beef cattle in Australia (AU) and the United States (US) were estimated using Restricted Maximum Likelihood. Five different uni- and bivariate models were used to fit both traits within each country. In Model 1, only a direct genetic effect (a) was fitted. In Models 2 and 3, a maternal genetic effect (m) was added. A genetic covariance between direct-maternal effects [cov(a, m)] was ignored (model 2) or included (model 3). Models 4 and 5 both m and maternal permanent environment effects (pe), were allowed from model 1, and assumed cov(a,m) in the same manner as model 2 and 3, respectively. When ignoring m effect, the direct heritability estimates were inflated and differed markedly from other models. The likelihood ratio test showed that model 5 was the best fit for both traits in the US while models 2 and 4 were the fittest for BW and WW in AU, respectively. The estimates of direct, maternal, total heritabilities, and maternal permanent environment variance of the full model in AU and in the US (in parentheses) were 0.46 (0.48), 0.09 (0.05), 0.42 (0.42) and 0.00 (0.06) for BW, and 0.23 (0.32), 0.16 (0.09), 0.24 (0.26) and 0.13 (0.10) for WW, respectively. After m and pe were fitted, the estimate of total heritability decreased slightly for BW in both countries and for WW in AU, but decreased re-markedly for WW in the US. Estimate of direct-maternal genetic correlation was moderately negative and tended to be more negative after pe was fitted for both traits in the US. The parameters estimates using bivariate analysis were not different to the results from univariate analysis. This analysis yielded additive and maternal genetics correlations between BW and WW. These estimates were positive and medium to high correlation, which were higher in AU than in the US. Correlation of estimated breeding values for direct additive and maternal genetics between the full model and others were high and close to unity. The differences of some parameters between both countries indicate that joint genetic evaluation might require genotype by environment interaction to be considered.