Supplementary Figure 2: Performance of expression prediction models in GTEx and TCGA datasets for genes with at least 10% correlation in GTEx data.

The x axis represents the prediction performance (R2) in the GTEx dataset (n = 67). The y axis represents the prediction performance in the TCGA dataset (n = 86). Each dot represents the expression prediction model for one gene. There is a trend that genes with high internal prediction performance in GTEx data also have high external prediction performance in TCGA data (Pearson's correlation coefficient: 0.55).