A method for supporting retrieval of articles on protein structure analysis considering users’ intention

BMC Bioinformatics

Research

Tomonobu Ozaki¹  Ryo Fujimoto²  Takenao Ohkawa³  Riku Kyogoku³ 

[1] Cybermedia Center, Osaka University, 565–0871, Yamadaoka, Suita, Japan;Graduate School of Engineering, Kobe University, 657–8501, Rokkodai, Nada, Kobe, Japan;Graduate School of System Informatics, Kobe University, 657–8501, Rokkodai, Nada, Kobe, Japan;

关键词: Gene Ontology;    Edge Weight;    MeSH Term;    Retrieval Result;    Related Article;   

DOI  :  10.1186/1471-2105-12-S1-S42

来源: Springer

【 摘 要 】

BackgroundIn recent years, information about protein structure and function is described in a large amount of articles. However, a naive full-text search by specific keywords often fails to find desired articles, because the articles involve the ambiguous and complicated concepts that cannot be described with uniform representation. For retrieving articles on protein structure and function, it is important to consider the relevance between structural and/or functional concepts by identifying the user’s intention.ResultsWe introduce a scheme of evaluating relevance between articles based on various biological databases and ontologies on structures and functions of proteins. The relevance, which is defined as a path length between concepts on hierarchies, is modified adaptively based on additional articles as a query in order to reflect the user’s intention. Also we implemented the retrieval system, in which the user can input some articles as a query and the related articles are retrieved and displayed on the 2D map.ConclusionsThe effectiveness of the proposed system was confirmed experimentally by having shown that the users can obtain easily highly related articles which reflect their intention.

【 授权许可】

CC BY   
© Kyogoku et al; licensee BioMed Central Ltd. 2011