MOESM2 of Joint learning improves protein abundance prediction in cancers

Additional file 2: Supplementary Tables. Table S1. The five-fold Pearson’s correlations of the generic, gene-specific and trans-tissue models. Table S2. The five-fold correlations of models using different numbers of features. Table S3. The weighting ratios to stack the generic, gene-specific and trans-tissue models and the corresponding prediction correlations in breast. Table S4. The weighting ratios to stack the generic, gene-specific and trans-tissue models and the corresponding prediction correlations in ovary. Table S5. The five-fold correlations of models using different numbers of samples. Table S6. The correlations and RMSEs across 32 overlapping ovarian cancer samples measured at both JHU and PNNL. Table S7. The correlations and RMSEs of our predictions on the held-out testing dataset of 82 ovarian cancer samples during the NCI-CPTAC DREAM challenge. Table S8. The feature importance of all genes in breast. Table S9. The feature importance of all genes in ovary. Table S10. The list of driver genes and their clusters in the breast functional network. Table S11. The list of driver genes and their clusters in the ovary functional network.