Interpreting the importance of transcript features in protein abundance prediction models

Supplementary Tables S1-S8.
Protein and mRNA levels correlate only moderately across samples. The availability of proteogenomics data sets with protein and transcript measurements from matching samples is providing new opportunities to assess the degree to which protein levels in a system can be predicted from mRNA information. We analyzed large proteogenomics datasets from the Clinical Proteomic Tumor Analysis Consortium (CPTAC) containing 13 637 proteins with matching transcriptome data from up to 958 tumor or normal adjacent tissue samples each, and trained models to predict protein abundance from the transcriptome of a sample using different algorithms and feature sets.