Predictive ability of sire weaning weight breeding values estimated across environments in terminal sire sheep breeds

Abstract

Data used for the genetic evaluation of terminal sire sheep breeds in Australia originates from a wide array of genotypes and environments. As a result there are large differences in the levels of production and therefore contemporary group means within the data. Differing variances within contemporary groups results in scale effects on the estimated breeding values (EBVs), as there is a tendency for standard deviations to increase with the mean level of expression, groups with a higher phenotypic mean often have more phenotypic variation and, without adjustment, animals from these groups have greater variation in their EBVs. Consequently, animals are likely to be selected from the more variable groups, especially if selection is intense (Hill, 1984). This heterogeneity in variances across contemporary groups results in EBVs that do not reliably predict progeny performance across the whole range of production environments, and this in turn leads to lower confidence in the use of breeding values across flocks where environments and management practices may differ. A relatively simple way of accounting for heterogeneity in variances across fixed effect levels is the transformation of observations to a proportion of their contemporary group mean, an approach regularly used in the Merino sheep industry, as was previously shown by Brown et al. (2005) for wool traits in Merino sheep. The aim of this work was to investigate the predictive ability of terminal sire EBVs for progeny performance in low, medium and high production environments in Australia using weaning weight using different data transformations.