Each animal in the base population was assigned a direct and indirect breeding value, drawn from N 0 0, σ A D 2 σ A DI σ A DI σ A I 2. The σ A D 2 and σ A I 2 were set to 1.00, and σ A DI was set to −0.50, 0.00 or 0.50.
The values for time were integer values drawn from a uniform distribution.
Simulation was initialized with concentration values drawn from a uniform distribution between 0.3 and 0.7.
To generate residual polygenic effects, breeding values of founder animals were drawn from a normal distribution N (0, σ u 2 ) with polygenic variance σ u 2 set to 30% of the total genetic variance.
Artificial selection was performed using the best linear unbiased prediction (BLUP) of breeding values obtained from an animal model.
The statistical properties of breeding values resulting from the mixed model approach still hold for a population under selection.
Estimated breeding values from routine breeding estimation were used for fertility traits.
GS is based on the prediction of breeding values from individual genotypes (estimated genomic breeding values, GEBV).
The EBV1 values were therefore used as the true breeding values, the standard against which other estimated breeding values from the cross-validation studies were compared.
More specifically, for each iteration from the model, we computed the average breeding values per generation of (i) actual estimated breeding values from the population and (ii) simulated breeding values under drift (rbv function in MCMCglmm).
