项目名称: 基于商空间理论的蛋白质相互作用研究
项目编号: No.61203290
项目类型: 青年科学基金项目
立项/批准年度: 2013
项目学科: 自动化学科
项目作者: 杜秀全
作者单位: 安徽大学
项目金额: 24万元
中文摘要: 蛋白质是生命体的主要基本物质,对蛋白质相互作用的深入研究不仅有助于理解细胞喝生物通路的功能变化,而且理解这些相互作用对各种疾病的发病机理和治疗具有积极的推进作用。目前存在大量的方法从不同的角度分析和预测蛋白质相互作用,但精度依然比较低。本研究引进一种新的理论分析方法,以公共数据库中的蛋白质相互作用数据为基础,利用商空间理论中的保假原理和保真原理将原始问题粒度化,再分别对各商空间下的预测模型采用粒度合成方法进行结果合成,最后将构造的商空间模型应用于蛋白质相互作用网络的构建。探讨不同特征下的蛋白质相互作用界面残基特性,设计一种新的覆盖聚类算法进行蛋白质相互作用位点的识别,并建立相应的网络在线预测系统,为生物学家进行相关的实验研究提供方向参考。
中文关键词: 蛋白质相互作用;商空间模型;特征选择;多示例;驱动变异
英文摘要: The protein is mainly basic material of body, for protein-protein interaction with in-depth study is not only helpful to science solution of cellular biological pathways and functional changes, but also the understanding of these interactions in a variety of disease pathogenesis and treatement with product pole propulsion action. At present, there are a lot of methods from different angles analysis and prediction of protein-protein interactions, but precision is still low. This research intends to introduce a new theoretical analysis method using public databases of protein-protein interactions. Based on the data, using false principle and true principle of the theory of quotient space in order to the original problem coarse graining. Respectively on analysis of the different size and the protein-protein interaction data to construct a corresponding granularity quotient space model. The prediction model using granular synthesis method to synthesis results, finally the structure of the quotient space model applied to protein interactions. Explore the different characteristics of the protein interaction interface residues, design a new covering clustering algorithm method for the identification of protein-protein interaction sites, and establish corresponding online prediction system for biologists, related exper
英文关键词: protein-protein interaction;quotient space model;feature selection;multi-instance;driver mutation