Including non-additive genetic effects in Bayesian methods for the prediction of genetic values based on genome-wide markers

Table 6 Comparison of empirical variances of predicted genetic values and genetic variance components estimated under LE*

	Simulation without epistasis
		23-QTL scenario			230-QTL scenario
Model		$σ_{a}^{2}$	$σ_{d}^{2}$	$σ_{epi}^{2}$	$σ_{a}^{2}$	$σ_{d}^{2}$	$σ_{epi}^{2}$
M1	empirical	0.743	0.035	-	0.711	0.163	-
	under LE	0.742	0.035	-	0.699	0.165	-
M2	empirical	0.749	0.038	0.030	0.805	0.338	0.278
	under LE	0.748	0.039	0.030	0.732	0.304	0.243
	Simulation with epistasis
		23-QTL scenario			230-QTL scenario
Model		$σ_{a}^{2}$	$σ_{d}^{2}$	$σ_{epi}^{2}$	$σ_{a}^{2}$	$σ_{d}^{2}$	$σ_{epi}^{2}$
M1	empirical	1.309	0.161	-	0.981	0.266	-
	under LE	1.310	0.161	-	0.920	0.267	-
M2	empirical	1.332	0.192	0.554	1.442	1.112	1.277
	under LE	1.338	0.193	0.559	1.277	0.910	1.047

* fBayesB was used in both QTL scenarios with 5 227 markers and H² = 0.5. Estimates were obtained as empirical variances of effect-specific genetic values predicted in the validation set (rows "empirical") or as genome-wide sum of locus-specific genetic variances which coincides with the assumption of LE (rows "under LE"). Variance components for each source of genetic variation: $σ_{a}^{2}$ additive genetic, $σ_{d}^{2}$ dominance, $σ_{epi}^{2}$ joint contribution of all epistatic effects. M1 includes additive and dominance effects, M2 includes additive, dominance and pairwise epistatic effects, LE linkage equilibrium.

ISSN: 2730-6844