Refining the Protein-Protein Interactome using Gene Expression Data

Proteins interact with other proteins within biological pathways, forming connected subgraphs in the protein-protein interactome (PPI). However, proteins are often involved in multiple biological pathways which complicates inference about interactions between proteins. Gene expression data is informative here since genes within a particular pathway tend to have more correlated expression patterns than genes from distinct pathways. We provide an algorithm that uses gene expression information to remove inter-pathway proteinprotein interactions, thereby simplifying the structure of the protein-protein interactome. This refined topology permits easier interpretation of multiple biological pathways simultaneously.