利用基因表达谱研究精神分裂症多变量协同机制及分类算法

项目名称： 利用基因表达谱研究精神分裂症多变量协同机制及分类算法

项目编号： No.61202288

项目类型： 青年科学基金项目

立项/批准年度： 2013

项目学科： 计算机科学学科

项目作者： 卢新国

作者单位： 湖南大学

项目金额： 25万元

中文摘要： 通过精神分裂症基因表达谱研究人脑所处的功能状态,揭示脑功能障碍形成的内在机制,为精神疾病诊断提供生物标记.目前研究主要将基因作为独立变量来影响大脑功能从而导致精神障碍.遗传学研究表明精神分裂症是多基因共同作用所致,本项目就精神分裂症多变量的协同认知机制展开研究:分别建立精神分裂症的基因协同表达模型和特征空间的协同表达模型,基于以上两种模型分别研究该疾病的多变量协同机制;通过该机制抽取多变量协同表达模式,并且构造多变量协同模式的样本似然度方法,提出适合多变量协同表达模式的精神分裂症患者和健康对照组的分类算法;利用SAM和信息熵的方法研究多变量协同表达模式发掘中的基因选择;利用基因本体论,基因集富集功能分析和代谢通路分析多变量协同表达模式,研究精神分裂症的内在脑神经机制.研究有望揭示精神分裂症基因之间的相互作用机制,获得该疾病的客观生物学标记,为疾病的诊治取得有意义的成果.

中文关键词： 基因共表达模型；基因选择；精神分裂症；权重共表达网络；

英文摘要： Brain function state can be acquired via analyzing gene expression profiles of schizophrenia patients. Then the intrinsic mechanism of brain cognitive impediment is discovered. And the diagnosis labes of schizophrenia are achieved. Recently in many works the genes are treated as independent variant to study the brain function and mechanism of schizophrenia. The studies based on inheritance show that schizophrenia patient is caused by the joint efforts of many genes. So this application will work on the multivariate co-expression mechanism and classification of schizophrenia patients from health controls using gene expression profiles. Gene based co-expression model and feature space based co-expression model are proposed separately, and then the multivariate co-expression mechanisms of schizophrenia are presented respectively using these two models; in each mechanism the multivariate co-expression modes are extracted, and likelihood function between multivariate co-expression modes from different brain function samples is constructed, then the multivariate co-expression mode based classification schizophrenia patients from health controls also is proposed; the method of significant analysis of Microarray and entropy based gene analysis are used to select the discriminating genes and then for extraction of multiv

英文关键词： gene co-expression model；gene selection；Schizophrenia；weighted gene co-expression network analysis；