Additional file 1: Table S1. Title: Complete description of the phenotypes and datasets used in this study. Table S2. Title: Cross-validated accuracies for all traits and datasets with the different kernels. Description: Two cross-validation schemes are used. Stratified cross-validation ensures that each validation sample is a representative sample of the complete dataset. Leave-one-cluster-out cross-validation focuses on prediction accuracies for closely related individuals with no close relatives in the training set.