TABLE 4

Cross-validation (CV) correlation between predicted and observed phenotypes, obtained in a 10-fold CV conducted for female flowering (FFL), male flowering (MFL), the MFL to FFL interval (ASI) of 284 maize lines genotyped for 1148 markers, and grain yield (GY) of 264 maize lines genotyped for 1135 markers

Modela
Trait–environmentM-RKHSM-BLBLUPb
MFL-WW0.6070.790c
MFL-SS0.6740.7780.464
FFL-WW0.5880.781c
FFL-SS0.6480.7740.521
ASI-WW0.5470.5130.469
ASI-SS0.5720.5170.481
GY-WW0.5140.5250.515
GY-SS0.4530.4150.442
  • Three models were fitted to each trait (FFL, MFL, ASI, and GY) and environment (SS, severe drought stress; WW, well watered) combination.

  • a Models were molecular marker (M) using reproducing kernel Hilbert space (M-RKHS) regression, molecular marker (M) regression model using the Bayesian LASSO (M-BL), and best linear unbiased predictor (BLUP) using marker genotypes.

  • b Values of genetic variances used to compute BLUP ranged from 0.000 to 0.319 for flowering, and from 0.017 to 0.206 for grain yield.

  • c BLUPs were not computed because the estimated genetic variances were negligible.