Automated PageRank-based Sentences Ranking to Identify Protein Relations from Literature

• Main Authors: Yong-Xi Liu, 劉詠熙
• Other Authors: Jung-Hsien Chiang
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/60729765870799735470

碩士 === 國立成功大學 === 資訊工程學系碩博士班 === 95 === Along with the improvement of information and computational techniques, increasing number of biomedical researches and literatures have been reported at the public databases such as PubMed. As for identifying protein-protein interactions, there have been some related databases manually evidence and extract interaction data from biomedical literatures. But they offer limited information and the process is time-consuming and expensive in labor power. To enhance the protein-protein interaction extraction process, we implemented an automated framework that combing hierarchical template-based sentence matching and PageRank-based sentence ranking approaches. Using this framework, we extract the interaction evidence sentences and their interaction relations. In this research, we implement a text-mining system to identify many important relations in KEGG pathway databases and discover a great number of novel relations that could potentially extend the existing protein interactions and pathways databases.