|Ferris, T - MICHIGAN STATE UNIV|
Submitted to: American Dairy Science Association Proceedings
Publication Type: Abstract Only
Publication Acceptance Date: July 26, 2000
Publication Date: N/A
Technical Abstract: Two of the methods to validate estimation of genetic trend that were established by the International Bull Evaluation Service (Uppsala, Sweden) are 1) comparison of trends estimated from bull evaluations based on first versus all parities and 2) analysis of daughter yield deviation (DYD) within bull by first calving year of daughters. For US Holstein bulls, annual trend was 95 kg of breeding value for milk for evaluations based on first parities and 113 kg for evaluations based on all parities. As expected, no trend in bull DYD was found for evaluations based on first parities; however, annual trend in bull DYD for milk was 6 kg for the 1980's and 20 kg for the 1990's for evaluations based on all parities. Standard deviations for milk yield within herd increased from 1960 through 1999 in lactation records used in USDA genetic evaluations. Although adjustments were implemented in 1991 in the USDA evaluation system for heterogeneous variance within herd across time, an annual trend of 12 kg of milk was found for herd-year variances for first parities from 1960 through 1996. Further, within-herd variances among lactations were found to be unequal; variances for later lactations increased compared with those for first lactations, particularly during the 1990s. Although incorrect adjustment for parity can bias estimates of genetic trend, an effect for parity has been included in the USDA genetic evaluation model. A large reduction (up to 15%) in variances for later parities was required to eliminate the differences in trend estimates between evaluations based on first and all parities.