Table 5 WHEAT4 data set. Average Pearson’s correlations between observed and predictive values (and their standard deviation in parentheses) for 3 methods for a single-environment model for 50 random partitions with 80% of the lines in the training set and 20% of the lines in the testing set. Methods GB, GK, and AK are the GBLUP, Gaussian Kernel and Arc-Cosine Kernel, respectively. The best predictive model for each environment is in boldface. The three GB, AK, and GK were applied to data from NIR1 (first derivative), NIR2 (second derivative), markers, pedigree and some combinations. The best predictive model for each type of data used is in boldface
Data usedGBAKGK
NIR10.349 (0.07)0.347 (0.07)0.354 (0.07)
NIR20.346 (0.07)0.367 (0.07)0.354 (0.07)
GENOMIC0.424 (0.07)0.456 (0.07)0.454 (0.07)
GENOMIC+NIR10.436 (0.07)0.462+ (0.07)0.456 (0.07)
GENOMIC+NIR20.435 (0.07)0.466+ (0.07)0.455 (0.07)
Pedigree0.396 (0.07)
GENOMIC+Pedigree0.437 (0.07)0.450 (0.07)0.454 (0.07)
Pedigree +NIR10.420 (0.07)0.413 (0.07)0.425 (0.07)
Pedigree +NIR20.421 (0.07)0.418 (0.07)0.421 (0.07)
GENOMIC+Pedigree+NIR10.448 (0.07)0.455 (0.07)0.462 (0.07)
GENEMIC+Pedigree+NIR20.448 (0.07)0.460 (0.07)0.459 (0.07)
  • + Significant at the 0.05 probability level of the t-test for the hypothesis that the average of the correlation of kernel AK is superior to the mean of the correlation of kernel GB.