项目名称: 基于多源信息融合的受体和抗菌肽分层多标签分类预测模型研究
项目编号: No.31260273
项目类型: 地区科学基金项目
立项/批准年度: 2013
项目学科: 生物科学
项目作者: 肖绚
作者单位: 景德镇陶瓷学院
项目金额: 50万元
中文摘要: 研究受体和抗菌肽的结构、功能和定位信息已成为当前生物学和药理学研究热点问题,但通过生物实验方法确定这些信息存在很多问题,如成本高、周期长,甚至有些当前的技术还无法实现。随着网络及信息技术的发展,使得获取不同尺度、不同层面的多源生物信息成为可能。本项目将融合蛋白质元胞自动机图(反映不同位置氨基酸的相互影响)、物理化学属性矩阵图(反映氨基酸之间的物理化学属性关系)、PSSM矩阵图(反映序列中氨基酸的遗传进化信息)、GO(Gene Ontology)和功能域等信息,比仅利用单信息源或非协调利用部分多源信息获取更精确和更稳健的性能,而蛋白质离散灰色模型、模糊K近邻、支持向量机等模式识别算法所提供的互补信息进一步提高融合决策的精度。项目还将设计出一种高效的多标签分类器,与现有ML-KNN、神经网络等算法进行集成最终实现准确可靠的预测抗菌肽和受体的结构和功能。此概念模型对指导新药设计具有应用前景。
中文关键词: 生物序列;药物靶标;多标签分类器;基因本体;信息融合
英文摘要: Studying the structure, function and localization information of acceptor and the antibacterial peptide has become the focus of current biological and pharmacological, but there are many problems to determine these information through biological experimental methods, such as high cost, long cycle, and even some current technology can not be achieved. However, with the development of network and information technology, it is possible to make access to different scales and different levels of multi-source biological information. The project coordinate the utilization of the protein cellular automaton image(reflecting the interaction of the different positions of amino acids), physical and chemical properties of the matrix(reflecting the relationship of physical and chemical properties between amino acids), PSSM matrix(reflecting the amino acid sequence phylogenetic information), GO(Gene Ontology) and functional domain information can get more accurate and more robust performance, compared with only a single source of information. And protein discrete gray model, fuzzy K-nearest neighbor, support vector machine and other pattern recognition algorithms to provide complementary information to further improve the accuracy of the integration of decision-making. The project will also design an efficient multi-label clas
英文关键词: Biologic sequence;Drug target;Multi-label classifier;Gene ontology;Information fusion