Alternate methods for estimating breeding values for faecal egg count data from merino studs across Australia

Abstract

Selection for resistance to internal parasites is of interest to many sheep breeders. However, the genetic evaluation of faecal egg count is problematic due to the variable levels of expression, as a result of interactions with the environment, the species of parasite, and its skewed distribution. Transformation, variance standardisation and adjustment for heterogeneous variances are used to overcome these problems. This study aims to identify which combination of techniques produces the most accurate estimated breeding values (EBVs). The EBVs from analysis of cube root transformed faecal egg count with (EBV_S) and without (EBV_3) variance standardisation are highly correlated (0.96). Furthermore the genetic correlation between these traits was also very high (0.95) indicating that genetically these traits are the same. EBVs estimated after homogenising the residual variances across groups produced EBVs less influenced by the level of variance in the raw data. Analysing faecal egg counts without variance standardisation did not significantly reduce the accuracy of the genetic evaluation. However, the EBVs need to be expressed on a scale that breeders can interpret.