Hybrid data mining for discovery of knowledge about protein functions

提出了从蛋白质三维结构数据、网络数据(如蛋白质-蛋白质相互作用数据和代谢途径)以及描述蛋白质结构分析结果的文本数据中进行混合数据挖掘的几种方法。基于所提出的数据挖掘方法，已经发展了在蛋白质功能知识发现中的应用，包括蛋白质功能位点的预测、从生物医学文章中提取信息、考虑用户角度的相似文章检索以及从生物功能的角度对生物进行分类。

Several methods for hybrid data mining from the various types of data such as the protein 3D structure data, the network data(e. g. protein-protein interaction data and metabolic pathways), and the text data describing the results of the protein structure analysis have been proposed. Applications to the knowledge discovery of protein functions, which include prediction of the protein functional sites, information extraction from the biomedical articles, retrieval of similar articles considering user's aspect, and classification of organisms from the viewpoint of the biological function, have been developed based on the proposed data mining methods.