患病-未患病同胞的关联回归分析与应用  

Regression Models and Application for Association Study Between Affected and Unaffected Siblings

作　　者：李彩霞[1] 黎培兴[1] 方淯靖[2] 方积乾[3] 

机构地区：[1]中山大学数学与计算科学学院,510275 [2]中山大学肿瘤防治中心 [3]中山大学公共卫生学院

出　　处：《中国卫生统计》2016年第1期49-51,共3页Chinese Journal of Health Statistics

基　　金：第44批教育部留学回国人员科研启动基金资助

摘　　要：目的当核心家系中无父母信息时,经典的传递不平衡检验(TDT)已不再适用,本文针对患病-未患病同胞家系数据,提出相应的关联分析方法,并校正协变量的影响。方法我们把家系中的受累同胞作为病例,未受累同胞作为匹配的对照,建立非均衡的条件logistic回归模型或分层Cox回归模型。我们把模型应用到96个华人原发性高血压核心家系的基因定位数据,分析校正协变量后ACE的I/D多态性、钠盐摄入水平与高血压的关系。结果 ACE与高血压的关联性受到钠盐摄入量高低的影响,ACE与钠盐摄入程度存在交互作用,较高的钠盐摄入患者其ACE的I/D多态性与高血压的关联性更强。结论条件logistic回归模型或分层Cox回归模型能校正协变量的影响,有效分析基因的主效应和基因与协变量的交互作用。Objective When the information of parents is not available in the nuclear family,classical transmission dis- equilibrium test（TDT） is no longer suitable. In this paper, association anaiysis methods,in which the covariates were adjusted, were developed for affected and unaffected siblings. Methods Using affected siblings as cases, and unaffected siblings as matched controls, an unbalanced conditional regression model and stratified Cox model were proposed. We used an example of gene mapping for essential hypertension（ESH） to illustrate our method, and showed the association between the I/D polymor- phism of the ACE gene, sodium intake level and ESH, after adjusting covariates. Results The results show that sodium intake modifies the association of ACE with ESH. There is interaction effect between sodium intake and ACE. The association between I/D polymorphism and ESH was stronger among persons with a higher sodium intake than in those with a lower sodium intake. Conclusion The gene＇s main effect and gene-covariate interaction were derived efficientiy after covariates adjusted by using conditional regression model or stratified Cox model.

关 键 词：遗传关联分析 条件logistic回归模型 分层cox回归模型 交互作用 

分 类 号：R195.1[医药卫生—卫生统计学]