Table 2 Cross-validation estimates of the quality of prediction of different models (average R2 and its SD)
ModelDMCDMYDtSILKPH
Population effects
 No QTL28.4 (SD 4.18)17.1 (SD 4.16)10.4 (SD 2.97)29.2 (SD 4.35)
Founder alleles
 Pop + GCA48.2 (SD 4.48)29.0 (SD 5.32)32.9 (SD 4.60)49.8 (SD 4.82)
 Pop + GCA + SCA47.4 (SD 4.58)27.3 (SD 5.13)32.1 (SD 4.81)48.3 (SD 4.78)
SNP within-group
 Pop + GCA48.8 (SD 4.33)30.3 (SD 4.29)39.7 (SD 5.96)46.9 (SD 5.26)
 Pop + GCA + SCA48.6 (SD 4.48)30.2 (SD 4.25)39.5 (SD 6.01)46.7 (SD 5.36)
Hybrid genotype
 Pop + Add48.4 (SD 4.21)28.9 (SD 4.82)35.6 (SD 5.36)44.9 (SD 4.96)
 Pop + Add + dominance48.2 (SD 4.23)28.7 (SD 4.78)35.3 (SD 5.48)44.6 (SD 5.00)
  • For the different traits (DMC, DMY, DtSILK, and PH), we considered models only including population effects or models including population effects and QTL effects considering different allele codings. For these later models, for each sampling, QTL detected in the whole data set had their effects in the training set tested following a backward procedure and only the significant QTL were considered in the prediction model. Predictions were based on GCA/additive effects only or on models considering also SCA/dominance effects significant at a 5% individual risk level. DMC, dry matter content; DMY, dry matter yield; DtSILK, female flowering date; PH, plant height; QTL, quantitative trait loci; Pop, population; GCA, General Combining Ability; SCA, Specific Combining Ability; Add, additivity.