Thus, the hyperparameter α ν in v l ∼ IG(α ν, α ν - 1 ) reflects the magnitude of residual heteroskedasticity in the responses of interest due to slaughter dates clusters.
Recall that the hyperparameter α ν defines the magnitude of non-systematic heterogeneity of residual variances for each of the selected traits as a function of slaughter dates clusters.
In turn, for loin muscle pH at 45 min postmortem, the posterior mean of α ν was below six, and the corresponding upper boundaries of its 95% HPD was approximately 10 in all cases, thus indicating milder cluster-driven residual heteroskedasticity for this trait, whereby the range of posterior means of v l was close to three-fold across all 33 slaughter dates clusters.
It is interesting to note that these top two individuals derived from one slaughter date cluster that had the largest posterior mean for the relative residual variance v l (data not shown).
For the real data application, predictive performance was assessed using a five-fold cross-validation (Daetwyler et al. 2013), whereby animals within each slaughter dates cluster were randomly assigned to five nonoverlapping data partitions or folds of nearly equal size (175 191 animals).
Stability was defined through three indicators 1) the number of slaughterhouses for which the cluster has been identified; 2) the number of years of slaughter for which the cluster has been identified; 3) the addition of the two previous indicators.
But there are steps that Washington could take to prevent Khartoum from allowing a new slaughter, of the two million civilians clustered in relief camps; for example, targeting the assets of Sudanese leaders, whose lifeblood, as John Prendergast, of the International Crisis Group, points out, flows from shell companies and fraudulent charities set up to siphon off the country's new oil wealth.
Number of years of slaughter (n=5) for which the cluster has been identified.
We modeled heterogeneous residual variances as presented in Equation (2), with τ and v = { v l } l = 1 33 specifying the fixed effects of sex and the random clustering effects of slaughter dates, respectively.
The general WGP model in Equation (1) was further expanded to incorporate clustering effects of slaughter dates d = { d q } q = 1 33 ∼ N (0, σ d 2 ) as well as polygenic effects u ∼N (0, A σ u 2 ), where A is a known pedigree-based additive relationship matrix.
