Figure 3.

Prediction performance of models trained in MESA populations and tested in METS

We predicted expression in METS using only gene models with R² > 0.01. The MESA population used to train each set of models is shown on the x axis, and the Spearman correlation between predicted and observed gene expression in METS is shown on the y axis. For each training population, only gene intersects of all prediction algorithms are shown in the plot. For example, in AFA, all gene intersects of EN, RF, SVR, and KNN are plotted.