Genetic evaluation of the length of productive life in Holstein cattle in the Czech Republic

Abstract

Together with fertility and type traits, length of productive life (LPL) is currently the most important functional trait in cattle breeding. The highest effect of longer productive life decreases the costs of replacement. Longer productive life also leads to a higher proportion of cows that are in later, high producing lactations (Vukasinovic et al., 1997). The length of productive life is defined as the number of days from the first calving to culling, death or censoring. In this context censoring means that the animal was still alive at the time of analysis or the loss of follow-up, i.e. the fate of an animal after a certain point in time was unknown. Functional longevity, which depends on the ability of the cow to avoid culling for involuntary reasons such as sterility or disease (Ducrocq, 1987), was used. The implementation of survival analysis techniques in statistical analysis and genetic evaluation for lon-gevity enabled to carry out appropriate analyses of longevity data (Vukasinovic et al., 1999). Survival analysis uses complete available information on both uncensored and censored records, provides a proper statistical treatment of censored records, and allows for changes in culling policy over time. Hazard of culling is defined as the product of the baseline hazard function and the function of variables affecting longevity. The results for fixed effects are expressed as relative culling rates, defined as the ratio between the estimated risk of culling under the influence of a certain environmental factor and the average risk which is usually set to one. Values higher than one indicate a higher culling risk associated with the environmental factor. Relative culling rate lower than one indicates a lower culling risk, i.e. increasing effect of the environmental factor on longevity (Vukasinovic et al., 2001). Boettcher et al. (1999) reported that the survival analysis also tended to give higher estimates of heritability than did the linear model when both estimates were expressed on the linear scale. ABSTRACT: Survival Kit V3.12 was used to analyse the length of productive life of cattle in the Czech Republic. The data set consisted of 230 028 registered Holstein cows. The model included the time-dependent effects parity × stage of lactation interaction, herd × year × season interaction, class of milk production within herd and year, breed within years and the time-independent effect of age at first calving and the random effect of sire. The highest risk of culling was found for cows at the beginning and at the end of the first lactation and at the end of any other lactation. The risk of culling decreased with parity. The risk of culling of cows assigned to the lowest milk production class was five times higher than that of cows assigned to the average milk production class. Risk of culling diminished with a decreasing percentage of Holstein breed. Cows younger at first calving showed a lower risk of culling. Breeding values for sires expressed as a risk ratio of their daughters were between 0.7 and 1.45. Estimated heritability of functional longevity was 0.025 on the log scale and 0.041 on the original scale.