Abstract
The objectives of this study were to determine the optimal genetic evaluation model by random regression for Somatic Cell Score (SCS) in first three lactations due to comparing additive genetic and permanent covariance structure on test day SCS with that of multi-traits analysis and to estimate 305 days heritabilities of SCS for Hokkaido’s Holstein cows. Multi-trait analysis by assuming different traits in yield at different stage of lactation and random regression test day models with multi-lactation were used in the analyses. In random regression analysis, variance components were estimated via Legendre polynomials (LP) in days of order 0（repeated）or 1 for additive genetic and order 1, 2, 3 or 4 for permanent environment. The criterions of model selection were bias, mean square error and correlations on sum of predicted additive genetic effects or sum of predicted permanent environment effects in 305 days compared with multi-traits results. The optimal genetic evaluation model of SCS was LP with order of 0 for additive genetic, and order of 3 for permanent environment effects. Heritabilities of 305-day SCS were 0.18, 0.19 and 0.20 for first, second and third lactation respectively.
