Table 4.

Comparison of mean predictive ability across 50 cycles of cross-validation for heading date, plant height and powdery mildew resistance according to genomic selection models having no marker as fixed effects with models that consider the addition of trait associated markers as fixed effects. Analyses used the Random training population selection method and different training population sizes. Combinations of diagnostic markers for loci associated with heading date (Ppd-D1, Vrn-A1, Vrn-B1) and plant height (Rht-B1, Rht-D1) were utilized, while for powdery mildew resistance the most significant SNP detected in GWAS for each validation cycle was utilized as the fixed effect

	Training population size
	TP050	TP100	TP150	TP200	TP250	TP300	TP350
Heading date
No fixed marker	0.46	0.53	0.56	0.61	0.63	0.67	0.68
Ppd-D1	0.42*	0.54	0.58	0.62	0.64	0.68	0.69
Vrn-A1	0.51**	0.58**	0.60**	0.63*	0.65*	0.68	0.69
Vrn-B1	0.49	0.56	0.59	0.63	0.64	0.67	0.68
Vrn-A1_Ppd-D1	0.51**	0.59**	0.62**	0.65**	0.67**	0.70**	0.70**
Vrn-A1_Vrn-B1_Ppd-D1	0.56**	0.61**	0.64**	0.67**	0.68**	0.71**	0.71**
Plant height
No fixed marker	0.49	0.53	0.55	0.58	0.60	0.61	0.63
Rht-B1	0.47	0.52	0.55	0.58	0.60	0.61	0.62
Rht-D1	0.56**	0.60**	0.62**	0.64**	0.66**	0.67**	0.68**
Rht-B1_Rht-D1	0.59**	0.64**	0.67**	0.69**	0.71**	0.72**	0.73**
Powdery mildew resistance
No fixed marker	0.36	0.44	0.48	0.51	0.52	0.53	0.55
Most significant SNP	0.42**	0.50**	0.53**	0.56**	0.56**	0.57**	0.60**

*, **Significantly different from the no fixed marker model at α = 0.05 and α = 0.01, respectively