基于信息增益理论的整体基因中基因互作挖掘方法  

A gene interaction mining method of the whole gene based on the information gain theory

作　　者：黄冬丽[1] 郭茂祖[1] 李晋[2] 刘晓燕[1] 王春宇[1] 

机构地区：[1]哈尔滨工业大学计算机科学与技术学院,哈尔滨150001 [2]哈尔滨医科大学生物信息科学与技术学院,哈尔滨150081

出　　处：《中国科技论文》2015年第2期186-191,共6页China Sciencepaper

基　　金：国家自然科学基金资助项目(61172098;61271346;91335112);高等学校博士学科点专项科研基金资助项目(20112302110040);中央高校基本科研业务费专项资金资助项目(HIT.KISTP.201418)

摘　　要：复杂疾病一般由多个基因共同作用发生,单个基因的效应微小,为了更好地研究基因互作对复杂疾病的影响,提出了一种基于基因的信息增益模型。信息增益在分类系统中指变量为分类带来信息的多少,带来的信息越多,该变量对分类越重要。该模型从一个整体基因的所有单核苷酸多态性位点(single nucleotide polymorphism,SNP)出发,采用病例-对照数据来检测基因互作对疾病的影响。由于基因是功能表达的最小单位,与基于SNP的交互作用分析方法相比,该模型更能从生物学的角度解释疾病的遗传机制。最后,采用模拟数据和类风湿性关节炎疾病的真实数据进行实验,并与基于SNP的熵模型以及基于基因的核典型相关分析模型(kernel canonical corelation based U statistic,KCCU)两种模型比较,结果均验证了该模型的有效性。Complex diseases generally are caused by the interaction of multiple genes,and the effect of a individual gene is small.In order to study the gene-gene interactions on complex diseases,agene-based information gain model is proposed.Information gain in classification system refers how much information a variable can bring.More information the variable brings,more important it is for the classification system.The model All SNPs of a gene are applied to detect gene interaction on disease,according to the case-control data.Since the gene is the smallest unit of the functional expression,compared with SNP-based interaction model,the model can explain the genetic mechanisms of disease from the perspective of biology.Finally,by using simulated data and real data of rheumatoid disease to conduct experiment,and the compared the gene-based information gain model with the SNP-based entropy model and gene-based KCCU model.The results verify the validity of the model.

关 键 词：复杂疾病 基因互作 整体基因 信息增益 

分 类 号：Q75[生物学—分子生物学] TP311.13[自动化与计算机技术—计算机软件与理论]