Differential response from selection for low birth weight versus high calving ease in beef cattle

The economic importance of calving ease is derived from the reduction in costs associated with dystocia. However, the genetic improvement of calving ease did and still does rely upon the downward selection for a trait with no direct economic relevance (i.e., birth weight). Given the antagonistic genetic relationship between calving ease and postnatal growth traits, such a strategy could result in production of lighter animals with uncertain gain in calving ease. Therefore, we hypothesized that direct selection for high calving ease would reduce performance losses associated with selection for low birth weight. Thus, the main objective of our study was to compare two selection approaches: 1.selection for high calving ease; and 2. selection for low birth weight. To evaluate these approaches, we used both simulated data and American Simmental Association field data. Another complicating factor was the approach to evaluation of calving ease with a threshold versus a linear model. The advantages of the threshold model over the linear model, in the analysis of ordered categorical traits, were investigated in the literature. Results are varied with some supporting and others discounting the superiority of the threshold model. Therefore, another goal of the current study was the predictive ability of the threshold and linear methodologies used in the genetic evaluation of calving ease as an example of ordered categorical traits.;The comparison of models predictive ability using threshold and linear or animal and sire approaches revealed that the threshold model outperformed the linear model. The highest predictive ability among all compared models was obtained from the threshold-linear sire model with calving ease fitted as a binary trait. The inclusion of linear trait(s) improved the prediction of categorical traits. Furthermore, the analysis of categorical traits with two continuous traits resulted in small differences between the threshold and linear models. The higher the number of categories, the better the linear model prediction; in contrast, the threshold-linear models showed better predictive ability when calving ease was fitted as a binary outcome. (Abstract shortened by UMI.).