Fig. 2: Prediction accuracy comparing the n = 31 core and the de-novo n = 100 core for prediction accuracies across models. | Communications Biology

Fig. 2: Prediction accuracy comparing the n = 31 core and the de-novo n = 100 core for prediction accuracies across models.

From: Environmental genomic selection to leverage polygenic local adaptation in barley landraces

Fig. 2

Cross-validation of genomic prediction. Four genomic prediction methods (RR-BLUP, G-BLUP with a Gaussian Kernel, G-BLUP with an exponential kernel, and BayesCπ) were evaluated using 6 cross-validation schemes. Prediction accuracy (r(PGE, y)) was estimated using 10-fold cross-validation with 50 replicates. The line shows the mean value across all runs and replicates, and the blue ribbon shows the mean value ± SD.

Back to article page