Paper: In Search of Protein Locations

ACL ID W11-0212
Title In Search of Protein Locations
Venue Workshop on Biomedical Natural Language Processing
Year 2011

We present a bootstrapping approach to infer new proteins, locations and protein-location pairs by combining UniProt seed protein- location pairs with dependency paths from a large collection of text. Of the top 20 system proposed protein-location pairs, 18 were in UniProt or supported by online evidence. In- terestingly, 3 of the top 20 locations identified by the system were in the UniProt description, but missing from the formal ontology.