Submitted to: World Congress of Genetics Applied in Livestock Production
Publication Type: Proceedings
Publication Acceptance Date: January 11, 1998
Publication Date: N/A
Interpretive Summary: Genetic evaluations based on lactation records have ignored differences in the shape of the lactation curve. Although models based on observations of daily yield can account for lactation shape and offer more precision, they use many more equations. A two-trait model that includes just yield and persistency of lactation was derived to account for lactation shape with far fewer equations. A "best prediction" method was used to condense information from many test days into optimal lactation measures of yield and persistency. These measures could be used by farmers who wish to continue to base their breeding, marketing, and culling decisions on lactation yields instead of individual test days.
Selection index procedures can condense observations of daily yield into estimates of lactation yield and persistency. Condensed data instead of test day data in an animal model might provide many of the same benefits with less computation. The best prediction method uses the phenotypic covariance matrix among daily yields to compute reliabilities and covariances of estimated yield and persistency for any data pattern. The measure of persistency is uncorrelated with yield if persistency coefficients are deviated from a central point of the lactation. Recent records containing both yield and persistency data and earlier records containing just yield data could be combined easily in genetic evaluations.