Abstract
The suitable model using TD animal models was examined to estimate the test day (TD) milk yield losses associated with elevated somatic cell score (SCS) in different stages of first and second lactations in Holstein cows. We used 200,095 TD milk and SCS records from 21,238 cows in their first lactation and 143,051 records from 15,281 cows in their second lactation. Stages of lactation were defined as 5 to 60 days in milk (DIM) (early stage), 61 to 150 (middle stage), 151 to 240 (late stage) and 241 to 305 (last stage). The data were analyzed within parity by using a single-trait random regression animal model. Class of classified SCS or linear to sextic regression to SCS in different stages were fitted as fixed SCS effect on milk in models, and the excellence of fit of these models were compared. The models applying quartic and more regressions to SCS were superior in the excellence of fit ; however, the effects of high SCS on milk yield were not estimated correctly. Therefore, it was suggested that class of classified SCS was preferable to estimate the effects of SCS on TD milk yield. Estimated milk losses associated with elevated SCS were greater in second lactation than those in first lactation and the losses increased as lactation stage progressed. These Estimates were significantly different between early/middle stages, late stage and last stage in both first and second lactation. Therefore, these stages of first and second lactation should be considered when estimating TD milk losses associated with SCS.
