Protein-protein relation extraction from biomedical abstracts

Recently, extraction systems have also used link grammar to identify interactions between proteins. In this chapter, the authors present a fully automated system to extract biomolecular events from biomedical abstracts. By semantically classifying each sentence to the class type of the event and then using high-coverage rules, BioEve extracts the participants of that event. The chapter explains in detail different classification approaches, and event extraction using a dependency parse tree of the sentence is explained here. It describes experiments with classification approaches, event extraction, and evaluation results for the BioNLP'09 shared task 1.