Multitrait analysis of growth traits for the optimization of breeding value prediction in Braunvieh cattle

Resumen

Currently, the genetic evaluations of growth traits (birth weight (BW), weaning weight (WW), and yearling weight (YW)) for the Mexican Braunvieh cattle are carried out in a univariate (for BW) and a bivariate (for WW and YW) models. Precision of genetic evaluations can be improved by a trivariate model. It was aimed to study bias in the univariate and bivariate evaluations due to the missing trait(s) in the analysis and the accuracy gain by the trivariate analysis. Pedigree and performance data were obtained from the Asociación Mexicana de Criadores de Ganado Suizo de Registro. After data edits, univariate, bivariate, and trivariate analyses were performed to make comparisons. A simple data pruning strategy was employed, considerably reducing the data size in the analyses. Animals excluded from the analyses were evaluated at a low computational cost from solutions of animals included in the analyses. The bivariate analysis showed biased WW and YW evaluations and genetic trends. The genetic trends underestimated young animals. Since the mid-1990s, all the traits showed a steady genetic progress. The bias was due to natural/artificial preselection on BW. The inclusion of BW in the trivariate analysis helped to consider the preselection information. The univariate BW evaluation and genetic trend were unbiased. Also, BW gained less accuracy from WW and YW than WW and YW from BW. Based on the results of this study, it is recommended the trivariate analysis of the traits with data pruning to lower the computational cost.