Comparative assessment of sire evaluation by univariate and bivariate animal model for estimation of breeding values of first lactation traits in HF cross cattle

Abstract

The aim of the present investigation was to study the superiority of bivariate over univariate sire evaluation. Data were collected on 1,988 first parity Karan Fries cows, spread over 31 years. The (co) variance components estimated by using average information restricted maximum likelihood (AIREML) were fitted into univariate and bivariate animal models for prediction of breeding values. Low heritability estimates were obtained for fertility traits ranging from 0.02 (FDPR) to 0.19 (AFC) indicating lesser role of additive gene action in fertility of dairy cattle. Comparative analysis revealed that the breeding values estimated using bivariate animal model had lower error variance and greater range in comparison to univariate animal models. The mean sire breeding values for production traits estimated by bivariate analysis ranged from 3055.50 to 3063.15 kg and were higher compared to the mean sire breeding values estimated by univariate animal model. The inclusion of fertility traits along with production traits improved the differentiating ability of bivariate animal model with respect to the production performance.