Table 1 Accuracy of genomic selection for grain yield (all environments, GYA; Wooster Ohio, GYW; Northwest Ohio, GYN), Fusarium Head Blight resistance (FHB), softness equivalence (SE), and flour yield (FY) using either all TP data (470 lines and 33,169 markers) or subsets of lines (n < 470) chosen for low genotype-by-environment interactions, and subsets of marker data (M2–M6) chosen based on significance criteria (see Materials and Methods)
TraitM1M2M3M4M5M6
No. markers for GYA, GYW, GYN3316929022524664293362
GYA, n = 4700.450.770.760.730.740.73
GYA, n = 4000.430.820.820.810.790.79
GYW, n = 4700.570.790.790.770.770.75
GYW, n = 4000.350.600.600.570.550.57
GYN, n = 4700.410.360.360.270.210.28
GYN, n = 4000.330.770.730.730.700.59
GYW to predict GYN, n = 470−0.07−0.10−0.09−0.08−0.06−0.08
GYW to predict GYN, n = 4000.130.310.320.330.350.37
GYN to predict GYW, n = 470−0.10−0.16−0.15−0.15−0.14−0.15
GYN to predict GYW, n = 4000.130.280.290.290.300.37
No. markers for FHB3316915561031286134
FHB, n = 4700.350.640.620.620.58
FHB, n = 4400.370.810.790.780.72
2010 FHB to predict 2011 FHB, n = 4700.130.170.170.170.21
2010 FHB to predict 2011 FHB, n = 4400.130.170.180.160.20
2011 FHB to predict 2010 FHB, n = 4700.150.320.310.320.36
2011 FHB to predict 2010 FHB, n = 4400.160.380.370.380.42
No. markers for SE3316916721133330151
SE, n = 4700.510.870.850.830.80
SE, n = 4470.510.890.870.850.82
2010 SE to predict 2011 SE, n = 4700.330.570.530.590.59
2010 SE to predict 2011 SE, n = 4470.320.580.530.590.50
2011 SE to predict 2010 SE, n = 4700.240.320.300.330.37
2011 SE to predict 2010 SE, n = 4470.240.320.310.330.26
No. markers for FY331691632968316166
FY, n = 4700.620.910.880.870.84
FY, n = 4470.620.910.880.870.84
2010 FY to predict 2011 FY, n = 4700.560.700.680.70.70
2010 FY to predict 2011 FY, n = 4470.590.740.720.740.74
2011 FY to predict 2010 FY, n = 4700.470.490.490.490.49
2011 FY to predict 2010 FY, n = 4470.480.500.500.50.50
  • Accuracy is shown for the trait itself using cross-validation and when using data from one set of environments to predict the phenotype from the other set of environments. Also shown is the number of markers in each marker subset for each trait.