Training-pop-p279_opti_diallel_STPGA.Rmd

This script takes as input files `geno-diallel_formated`, `geno-p279_formated` and `origin_CC279_UTF8.tsv`. It applies training set optimization with three methods (CDmean, PEVmean and mean additive relationship), for prediction both inthe whole half-diallel population and within each cross of the half-diallel. As a control, it also applies random sampling of cultivars from the diversity panel. The optimized training sets (i.e., subsets of the diversity panel) are saved in three files: `training-pop-p279_non-optimized_STPGA-varsize.Rdata` (control random sampling), `training-pop-p279_optimized-diall-all_STPGA-varsize.Rdata` (optimization for prediction in whole half-diallel), and `training-pop-p279_optimized-diall-cross_STPGA-varsize.Rdata` (optimization for prediction within each cross).