Search for a command to run...
Automatic milking systems record an enormous amount of data on milk yield and the cow itself. These type of big data are expected to contain indicators for health and resilience of cows. In this study, the aim was to define and estimate heritabilities for traits related with fluctuations in daily milk yield and to estimate genetic correlations with existing functional traits, such as udder health, fertility, claw health, ketosis, and longevity. We used daily milk yield records from automatic milking systems of 67,025 lactations in the first parity from 498 herds in the Netherlands. We defined 3 traits related to the number of drops in milk yield using Student t-tests based on either a rolling average (drop rolling average) or a regression (drop regression) and the natural logarithm of the within-cow variance of milk yield (LnVar). Average milk yield was added to investigate the relationships between milk yield and these new traits. ASReml was used to estimate heritabilities, breeding values (EBV), and genetic correlations among these new traits and average milk yield. Approximate genetic correlations were calculated using correlations between EBV of the new traits and existing EBV for health and functional traits correcting for nonunity reliabilities using the Calo method. Partial genetic correlations controlling for persistency and average milk yield and relative contributions to reliability were calculated to investigate whether the new traits add new information to predict fertility, health, and longevity. Heritabilities were 0.08 for drop rolling average, 0.06 for drop regression, and 0.10 for LnVar. Approximate genetic correlations between the new traits and the existing health traits differed quite a bit, with the strongest correlations (-0.29 to -0.52) between LnVar and udder health, ketosis, persistency, and longevity. This study shows that fluctuations in daily milk yield are heritable and that the variance of milk production is best among the 3 fluctuations traits tested to predict udder health, ketosis, and longevity. Using the residual variance of milk production instead of the raw variance is expected to further improve the trait to breed healthy, resilient, and long-lasting dairy cows.