Extracting Protein-Protein Interaction from.pdf

Biomedical Text Using Additional Shallow ParsingInformationHuanhuan YuLonghua Qian*Guodong ZhouQiaoming ZhuJiangsu Provincial Key Lab for Computer Information Processing TechnologySchool of Computer Science and TechnologySoochow UniversitySuzhou, ChinaAbstract—This paper explores protein-protein interactionextraction from biomedical literature using Support VectorMachines (SVM). Besides common lexical features, variousoverlap features and base phrase chunking information are usedto improve the performance. Evaluation on the AIMed corpusshows that our feature-based method achieves very encouragingperformances of 68.6 and 51.0 in F-measure with 10-fold pair- wise cross-validation and 10-fold document-wise cross-validationrespectively, which are comparable with other state-of-the-artfeature-based methods.Keywords-Protein-Protein Interaction; SVM; Shallow ParsingInformationI.INTRODUCTIONProtein-Protein Interaction (PPI) information frombiomedical literature plays a critical role in building proteinknowledge networks, predicting protein functions anddesigning new drugs. However, manual collection of relevantPPI information from thousands of new research papersincrementally published every day is so time-consuming thatautomatic extraction approaches with the help of NaturalLanguage Processing (NLP) techniques become necessary.Automatic PPI extraction from scientific literature is a taskof significant interest in the BioNLP field during recent years.The most commonly addressed problem has been the detectionof PPI information, where the system identifies which proteinpairs in a sentence have a biologically relevant relationship. Forexample, the sentence “TR6 specifically binds Fas ligand.”asserts an interaction relationship between the two proteins TR6