Alternative strategies for adjusting phenotype for environmental effects

Abstract

Performance data from the Lincoln University Dorset Down sheep flock for the years 1987-1991 were analyzed to estimate environmental effects. Several methods for estimating and correcting for significant effects were compared. This flock is involved in a selection programme to improve lean carcass weight while reducing the total carcass fat weight. Live weight, muscle depth and fat depth are used as the selection criteria to make gains in the objective traits. To achieve accurate estimates for each of these criteria, a correction method has been employed that adjusts for birth rank and date of birth within year-sex-line. However some obvious discrepancies have been found in this method because it did not discriminate between significant and non-significant effects. For this reason alternative approaches to this method were assessed in this study. For all three traits there were no significant interactions between year and other fixed (environmental) effects. For fat depth, but not liveweight or muscle depth, there were significant interactions between sex and birth rank, and sex and date of birth. Corrections for environmental effects (br, aod, dob) made from estimates across the whole dataset (n=1395) were similar to those obtained from the sex subsets (male n=695 female n=700). However, within year-sex-line data subsets gave more variable results due to greater errors in estimation associated with the small size of these data sets. It was concluded that standard correction factors could be defined for use across years, but that these would need to be separately estimated for ewes and rams.