Calculating cow and daughter yield deviations and partitioning of genetic evaluations under a random regression model

The equations for calculating cow yield deviations (YD) and partitioning genetic evaluations when using a random regression model (RRM) were presented. Calculations were illustrated using results from a RRM for the evaluation of test day (TD) somatic cell count (SCC) in the first three lactations of 1,287,930 Holstein/Friesian heifers. Random regressions for animal evaluations were partitioned into contributions from parent average, YD and progeny. These vectors of contributions from parent average, YD and progeny are vectors of regressions, therefore the relative contribution from these different sources of information can be computed for any day in milk. The equations for computing daughter yield deviations (DYD) for bulls were presented for the RRM. For bulls with almost 50 daughters, the correlation between DYDs and genetic evaluations was 0.77 for the RRM compared with 0.48 using an animal model (AM). The higher correlation may reflect a better correction of environmental factors in the RRM. The correlation increased to 0.97 with the RRM and 0.96 with the AM for bulls with 51-99 daughters. However, with 200 daughters or more, the correlations were 0.99 for both models. (C) 2003 Elsevier B.V. All rights reserved.