Technical note: Prediction of liveweight from linear conformation traits in dairy cattle

The objectives of this study were to derive phenotypic and genetic prediction equations of liveweight from linear conformation traits, and estimate genetic and phenotypic parameters for these traits. Data pertained to 2,728 conformation and liveweight records of 613 cows in 1,529 lactations. Cows were raised at the Scottish Agricultural College research station and had calved between 2002 and 2010. Fifteen linear conformation traits were considered as predictors. To validate phenotypic predictions, the data set was randomly split into independent reference and validation subsets. Reference subsets were used to derive prediction equations with the use of a mixed model. Comparisons between predicted and actual liveweight in the validation subsets indicated that stature, chest width, body depth, and angularity could be used to derive phenotypic predictions of liveweight. Accuracy of these predictions was better for first-lactation than for all-lactation liveweight data. Significant genetic correlations between liveweight and the 4 predictor traits ranged from 0.49 to 0.76, and phenotypic correlations were 0.33 to 0.56. Estimated genetic (co)variances were used to develop prediction equations of animal genetic merit for liveweight from routinely calculated genetic evaluations for conformation traits.