Abstract #286

# 286
Breeding for resilience in dairy cows using daily milk yield recording.
M. Poppe*1, H. Mulder1, R. Veerkamp1, 1Wageningen University & Research, Wageningen, the Netherlands.

Cows are constantly subject to environmental challenges, such as pathogens and heat waves. The ability to cope with such perturbations is called resilience. The objective of this research was to study the use of variability in daily milk yield resulting from perturbations as an indicator of resilience in dairy cows using a genetic analysis. Daily milk yield of 198,754 first-parity cows, recorded by automatic milking systems, were studied. To avoid influence of the shape of the lactation curve on the variability, first the general trend in milk yield was removed by fitting lactation curves for each cow. Four different methods were investigated: moving average, moving median, Wilmink curve and quantile regression. Next, the variability of the residuals was quantified by taking the log-transformed variance (LnVar). A lower LnVar would indicate less fluctuations and thus a more resilient cow. Heritabilities of LnVar, genetic correlations between LnVar based on different curve fitting methods, and genetic correlations with milk yield level were estimated using univariate and bivariate analyses in ASReml. Also, genetic correlations between de-regressed breeding values for LnVar and de-regressed breeding values for functional traits were estimated. The genetic analysis showed that LnVar had a moderate heritability (0.20 to 0.24) and that the genetic correlations between LnVar based on the different curve fitting methods were > 0.94. Because of large genetic correlations with average milk yield level (0.75 to 0.79), the correlations with functional traits were converted to partial genetic correlations, corrected for the correlations with average milk yield. LnVar had negative partial genetic correlations with udder health (−0.22 to −0.33), ketosis resistance (−0.27 to −0.30), and longevity (−0.29 to −0.34). This confirms the expectation that a lower variability in milk yield indicates a better resilience. For all functional traits, LnVar based on quantile regression gave the largest correlation, although the differences were small. Concluding, variance in daily milk yield is a promising resilience indicator that can be used on a large scale to breed resilient cows.

Key Words: resilience, automatic milking system (AMS), health