Investigation of the Environment Sensitivity of Merino Sheep ASBVs for Various Traits.

Waters, Dominic
University of New England


The objective of this thesis was to explore genotype by environment interactions (G×E) and macro-environmental sensitivity (macro-ES) in a multi-breed sheep population using multi- trait and reaction norm models. Post-weaning weight (PWWT), post-weaning growth rate (PWGR), post-weaning scanned fat depth (PCF), carcase weight (CWT) and intra-muscular fat (IMF) were investigated. The number of sheep analysed ranged from 28,860 for PWWT to 14,969 for IMF. Environment was defined by the best linear unbiased estimation (BLUE) of contemporary group effects in an animal model, using growth (PWGR) as the response variable. Each trait was first analysed using a univariate model to observe the estimation of genetic parameters under the assumption of no G×E. A multi-trait model was then used that considered performance in low, average and high growth environments as separate traits. Reaction norms were also used for each trait to estimate macro-ES as well as provide an alternative method for estimation of G×E. In the multi-trait approach, genetic correlations were significantly different from unity for all traits across all environments except for IMF.

Correlations ranged from 0.61 to 0.72 for PWGR, PCF and CWT, 0.77 to 0.86 for PWWT and 0.91 to 0.99 for IMF. Scaling effects were only significant for PCF. Overall, genetic parameters estimated using reaction norms were comparable to those estimated using multi- trait analysis, although the pattern of correlations between environments was different. Genetic variation in slope was significant for all traits except IMF. Quadratic polynomials were significant for PWWT, indicating that sensitivity was dependent on the level of the growth environment. Where estimated, correlations between the intercept and slope were moderately positive, indicating that current selection practices in Australia will increase the sensitivity of sheep for important production traits. The results of this thesis should be considered in the estimation of Australian sheep breeding values and highlights the potential for breeding sheep that are less sensitive to different environments.