Alternatives for modelling of traits within the calving traits complex

Abstract

Difficult calvings and losses of calf at birth are important factors for the welfare of the cow and the calf and also have severe economic consequences. For a long time in dairy cattle breeding, both characteristics have been regarded as separate although correlated traits. Scoring of calving ease bears the intrinsic problem of leading to a highly skewed distribution of observations across score classes. Stillbirth traditionally is considered as a well-defined trait. While this is true for the outcome of a calving, physiologically stillbirth has very different reasons and thus is needed to be re-defined in a more precise way. In the present study, using data from 81,419 calvings, it is attempted to find a new approach with consideration to the problems mentioned. Firstly, calving ease is considered as a binary trait since physiologically it makes more sense to differentiate the calvings as ‘normal’ or ’easy’ vs. ‘difficult’ or ’heavy’. Secondly, these two new alternative definitions of calving ease are matched with binary codes for stillbirth so that e.g. easy calvings resulting in a dead calf can be separated from heavy calvings also yielding a dead calf. Genetic parameters for a standard approach using a sire-mgs-model are compared with those for the newly defined traits. It can be shown that the magnitude and sign of the correlation between direct genetic and maternal genetic effects is depending on the definition of the respective trait. The approach is considered as useful for a supplementary analysis, especially when genomic data are jointly analysed in the form of a genome-wide association study.