Table 2.

Average normalized root mean squared error of prediction (NRMSE) and average Pearson’s correlation (Cor) in a 5-fold cross-validation strategy when predicting each one of the five traits (Yield, Maturity, Height, Heading and Germination) with GBLUP and DL models using E+G and E+G+EG as predictors.

Model	Trait	Predictor	NRMSE (SD)	Cor (SD)
GBLUP	Yield	E+G	0.0932(0.0008)	0.9203(0.0024)
GBLUP	Yield	E+G+GE	0.0908(0.001)	0.9245(0.0023)
GBLUP	Maturity	E+G	0.0259(0.0002)	0.9181(0.001)
GBLUP	Maturity	E+G+GE	0.0256(0.0002)	0.9199(0.0008)
GBLUP	Height	E+G	0.0536(0.0006)	0.7674(0.0091)
GBLUP	Height	E+G+GE	0.0532(0.0006)	0.7711(0.0093)
GBLUP	Heading	E+G	0.0405(0.0006)	0.8683(0.006)
GBLUP	Heading	E+G+GE	0.0399(0.0006)	0.8725(0.0058)
GBLUP	Germination	E+G	0.082(0.0024)	0.5721(0.0141)
GBLUP	Germination	E+G+GE	0.082(0.0025)	0.5727(0.0142)
DL	Yield	E+G	0.094(0.0011)	0.9189(0.0032)
DL	Yield	E+G+GE	0.0923(0.0014)	0.922(0.0029)
DL	Maturity	E+G	0.0249(0.0002)	0.9249(0.0011)
DL	Maturity	E+G+GE	0.0252(0.0003)	0.9229(0.0012)
DL	Height	E+G	0.0545(0.0013)	0.7588(0.0163)
DL	Height	E+G+GE	0.0541(0.0006)	0.7628(0.0088)
DL	Heading	E+G	0.0376(0.0011)	0.8885(0.008)
DL	Heading	E+G+GE	0.0389(0.0008)	0.8798(0.0062)
DL	Germination	E+G	0.0823(0.0025)	0.5689(0.0166)
DL	Germination	E+G+GE	0.0825(0.0024)	0.5685(0.016)

SD represents the standard deviation of the metric across the folds.