Top 50 proteins whose abundance is under substantial influence from non-cognate transcripts.

Columns 1 and 2: gene names. Columns 3 show the representative CORUM complex the gene of interest belongs to. Columns 4 and 5 denote the increase in prediction performance between elastic net single feature (self transcript) vs. CORUM feature sets. Column 6 shows the number of transcripts used to predict the protein level of the gene of interest in the CORUM feature set. Column 7 shows the top trans-locus contributor to the protein level of the gene of interest, ranked by absolute coefficients in the elastic net model. Proteins whose own transcripts are the top predictors are marked with (self). Column 8 denotes the number of MSigDB C2 CGP (chemical and genetic perturbation) gene sets in which the gene appears. Column 9 denotes the top significantly associated Disease Ontology term with the gene of interest in the literature.