MetadataShow full item record
Each day, scientists publish numerous articles in the area of biology and biomedicine. In this thesis, we describe a text mining project on automatic extraction of mutation impact information from scientific literature. The system identifies and predicts impacts of mutations in protein kinases on various properties of these important proteins. Our system performs sentence level mutation impact detection using grammatical dependencies present in the sentence. The SPARQL query language is used to capture and detect common grammatical patterns present in sentences describing mutation impacts. A curatorial system is used to verify the impact predictions created by our system. Furthermore, we have developed a semi-automatic module to detect new grammatical patterns, based on the feedback from curators, in order to improve the performance of the system. Once all mutation impacts have been identified, we intend to populate the ProKinO Ontology with the mutation impact data obtained from our system.