Table 2 Accounting for population structure
TraitTRN20-TST8010% Diversity Core20% Diversity Core10% Prediction Core20% Prediction Core
Mexican landraces
 Plant height (PHT)0.407 (0.006)0.3590.4120.3530.405
 Thousand-kernel weight (TKW)0.677 (0.007)0.6440.6540.6520.663
 Test weight (TW)0.498 (0.008)0.4570.4780.4620.497
 Grain hardness (GH)0.458 (0.008)0.4040.4500.4200.458
 Grain protein (GP)0.516 (0.009)0.4710.4970.4610.512
 SDS sedimentation (SDS)0.571 (0.007)0.5420.5390.45310.553
 Grain yield per square meter (GYSM)0.460 (0.006)0.4340.4510.4220.451
Iranian landraces
 Plant height (PHT)0.166 (0.027)0.1120.1820.1410.154
 Thousand-kernel weight (TKW)0.519 (0.017)0.4630.4680.4450.475
 Test weight (TW)0.437 (0.020)0.3920.3990.3910.379
 Grain hardness (GH)0.528 (0.017)0.4470.5200.3860.463
 Grain protein (GP)0.412 (0.023)0.4170.4080.3850.386
 Grain length (GL)0.662 (0.016)0.5930.6470.6120.628
 Grain width (GW)0.502 (0.019)0.4170.4750.4190.443
 SDS sedimentation (SDS)0.390 (0.021)0.3520.3770.3050.369
  • Mean correlation between predicted and observed values across 30 random cross-validation partitions (SD in parentheses) for a training set of 20% (TRN20) and a testing set of 80% (TST80), of the total Mexican and Iranian collections for several traits measured in a single environment using the GBLUP model. Correlation using 10% diversity and prediction cores, and 20% diversity and prediction cores as training sets to predict the remaining 90% and 80% of the collections.