Figure 6.
Average prediction correlations in testing sets from 100 random replications of across-group GBLUP (A-GBLUP), multigroup GBLUP (MG-GBLUP), and within-group GBLUP (W-GBLUP) evaluated within each of the four rice subpopulations [AUS, Indica (IND), Temperate Japonica (TEJ), Tropical Japonica (TRJ)]. Letters above bars indicate significant differences between models’ prediction correlations from paired t-tests evaluated within each subpopulation and trait, where estimates with the same letter are not significantly different (P < 0.05). (A) Flowering time. (B) Plant height. (C) Panicle length.