What genetic group structure to fit?: A Bayesian approach applied to yearling worm egg count data in Merino sheep

Abstract

Genetic groups need to be fitted in a genetic evaluation model to accommodate animals with unknown parents that come from a wide variety of sources. Different genetic grouping strategies were investigated for worm egg count data extracted from Sheep Genetics Australia's Merino database. A Bayesian approach was implemented that tested whether the genetic group variance was significantly different from zero. Four genetic grouping strategies were compared, 1) grouping on average fibre diameter, 2) groups by flock, 3) groups by flock-period, and 4) groups by flock-year. Two additional strategies were grouping strategies 3 and 4 with an assumed autocorrelation structure between genetic groups within flock.