Literature-based automated reconstruction, expansion, and refinement of the TGF-β superfamily ligand-receptor network.

The TGF-β pathway transduces a variety of extracellular signals into intracellular responses that control multiple cellular processes, including cell growth, apoptosis, and differentiation. It encompasses 33 ligands that interact with 7 type II receptors and 5 type I receptors at the plasma membrane to potentially form 1,155 ligand-receptor complexes in mammalian cells. Retrieving the information of the complexes that are actually formed from reading the literature might be tedious and prone to missing links. Here, we have developed an automated literature-mining procedure to obtain the interactions of the TGF-β ligand-receptor network. By querying the Information Hyperlinked over Proteins (iHOP) online service and processing the results, we were able to find pairwise interactions between ligands and receptors that allowed us to build the network automatically from the literature. Comparison with available published review papers indicates that this method is able to automatically reconstruct and expand the TGF-β superfamily ligand-receptor network. Retrieving and parsing the full text of the manuscripts containing the interactions allowed us to refine the network interactions for specific cell lines.