A novel method for genomic-enabled prediction of cultivars in new environments

Introduction Genomic selection (GS) has gained global importance due to its potential accelerate genetic progress and improve the efficiency of breeding programs. Objectives research In this we proposed a method prediction accuracy tested lines in new (untested) environments. Method-1 The trained model with modified response variable (a difference variables) that decreases lack non-stationary distribution between training testing improved accuracy. Comparing conventional We compared genomic best linear unbiased (GBLUP) (M1) including (or not) genotype × environment interaction (GE) (M1_GE; M1_NO_GE) versus (M2) on several data sets. Results discussion gain M2, M1_GE, M1_NO_GE terms Pearson´s correlation was at least 4.3%, while percentage top-yielding captured when selected 10% (Best10) 20% (Best20) 19.5%, Normalized Root Mean Squared Error (NRMSE) 42.29%.