Genetic improvement of longevity in dairy cows

Abstract

Improving longevity helps to increase the profit of the farmer, and it is seen as an important measure of improved animal welfare and sustainability of the sector. Breeding values for longevity have been published since in 1999 in the Netherlands. For AI-companies and farmers it is necessary that breeding values are accurately estimated and will remain stable for the rest of life. However, current breeding values for longevity of bulls seem to fluctuate more than expected. The main aim of this thesis was to revisit the genetics of longevity and develop a genetic evaluation model for longevity, where breeding values reflect the true breeding value quicker during early life and therefore breeding values become more stable. Genetic parameters were estimated for survival up to 72 months after first calving with a random regression model (RRM). Survival rates were higher in early life than later in life (99 vs. 95%). Survival was genetically not the same trait across the entire lifespan, because genetic correlations differ from unity between different time intervals, especially when intervals were further apart. Survival in the first year after first calving was investigated more in depth. Survival of heifers has improved considerably in the past 25 years, initially due to the focus on a high milk production. More recently, the importance of a high milk production for survival has been reduced. Therefore functional survival was defined as survival adjusted for within-herd production level. For survival the optimum age at first calving was around 24 months, whereas for functional survival calving before 24 months resulted in a higher survival. Over years, genetic correlations between survival in different 5-yr intervals were below unity, whereas for functional survival genetic correlations did not indicate that survival changed over years. This suggested that a genetic evaluation using historical data should analyze functional survival rather than survival. A new genetic evaluation system for longevity was developed based on a RRM analyzing functional survival. Based on the correlation between the first breeding value of a bull and his later breeding values, the ranking of bulls was shown to be more stable for RRM than the current genetic evaluation. Bias in breeding value was observed, mainly for bulls with a large proportion of living daughters. Adjusting for within-herd production level reduced this bias in the breeding values greatly. Before implementing this new model for genetic evaluation, the cause of this bias needs to be further investigated.