单核苷酸多态性数据的卡方及Logistic回归分析在SPSS统计软件中的实现  

Chi square and Logistic regression analysis of SNP data with SPSS 11.0

作　　者：韩瑞刚[1] 冯凯[1] 刘保池[2] 郑燕华[1] 路浩军[1] 张铭[1] 张建中[1] 

机构地区：[1]解放军第306医院,北京市100101 [2]河南省食管癌重点开放实验室

出　　处：《总装备部医学学报》2008年第2期63-65,共3页Medical Journal of General Equipment Headquarters

摘　　要：目的用SPSS11·0(statistical package fou social sciences)分析软件对贲门癌抑癌基因Ras相关区域家族1A多态性资料进行卡方分析,并用logistic回归分析计算不同基因型的优势比(odds ratio),以期为采用二分类因变量实验设计模型的临床与基础科研工作者提供可借鉴的统计学分析方法。方法应用SPSS11·0中二分类Logistic回归分析方法对病例-对照研究资料进行二值多元(非条件)Logistic回归分析。用单因素分析对自变量进行筛选,再对有显著性意义的自变量做多因素Logistic回归分析。结果可以得到影响患病与否的自变量,并可得到相应修正后的危险度。结论遗传学研究中的基因单核苷酸多态性(SNP)数据,属于因变量为分类变量的多变量资料,可归结于统计中的病例-对照研究资料,可以用Logistic回归分析探究其多态性意义,同时还可进行不同基因型危险度分析。应用SPSS统计软件易于实现该类数据的统计分析。Objective To introduce the Chi square method of analyzing the SNP data of Ras association domain family protein 1 gene in gastric cardia tumors and Logistic regression analysis of the odds ratio (OR) in different genotypes with SPSS 11.0, and then to offer a reference statistical method to researchers who take the design of binary dependents measures. Methods Using binary Logistic regression analysis in SPSS we studied the case-control data. Firstly, by single factor analysis all independents were screened, and then corrected-independents selected and analyzed by multinomial Logistic regression. Results Using the above methods, we identified the independents contributing to the disease, and the OR adjusted by some parameters. Conclusions SNP data in genetic research belongs to case-control data, whose dependents are binary variables and independents are multi variables. The SNP data are suitable for variance analysis of binary Logistic regression to evaluate the polymorphism significance and the OR of genotype. SPSS statistical package is available to fulfill this process.

关 键 词：LOGISTIC回归分析 危险度 SPSS统计分析软件 

分 类 号：Q75[生物学—分子生物学]