Figure 3.
Comparison of the evaluated methods in prediction of mouse phenotypes. Each dot represents the difference between the prediction performance of an evaluated method and AMB across one of the phenotypes, measured according to RMSE. Also shown is the average number of regions selected by each method across the training folds (rounded to the closest integer). MKLMM-Adapt outperformed AMB and GBLUP across 100 and 86 phenotypes, respectively.
