基于函数型Cox比例风险模型的基因–基因交互效应分析及在眼病中的应用  

Analysis of Gene-Gene Interaction Effects Based on Functional Cox Model and Its Application in Ophthalmopathy

作　　者：郭诗雨 李运明 郑海涛 

机构地区：[1]西南交通大学数学学院统计系,四川 成都 [2]西部战区总医院医疗保障中心,四川 成都

出　　处：《理论数学》2024年第2期817-827,共11页Pure Mathematics

摘　　要：疾病关联性研究存在大量的基因与基因的交互效应(gene-gene interaction)和基因与环境因素的交互效应(gene-environment interaction)分析,以上交互效应对个体化诊疗具有极为重要的参考价值。针对基因与基因交互效应,本文提出了一种具有函数交互效应的Cox比例风险模型。该方法将基因的多个单核苷酸多态性(SNP)之间的交互效应进行函数化处理,大大降低了待估参数的维数。基因–基因的交互作用的假设检验采用似然比(LRT)检验统计量。经模拟研究表明,所提方法能够较好地控制第I类错误率,功效也比较高。实例分析表明,利用所提出的方法能够有效地检测出与老年黄斑变性和白内障(AREDS)相关联的基因–基因交互作用。Gene-Gene interaction and Gene-Environment interaction have been widely used in disease associa-tion analysis. In particular, the interaction effect of personalized medicine has a very important re-search value. In this paper, a Cox proportional hazards model with functional interaction effect is proposed, which mainly studies Gene-Gene interaction effect. The interaction effects between multi-ple Single-nucleotide polymorphism of genes (SNPs) are functionally processed, which greatly reduces the dimension of the parameters to be estimated. The likelihood ratio (LRT) test was used to test for Gene-Gene interaction. A large number of simulation studies show that the proposed method can control the Type I error rate better, and the Power is also high. The proposed method can effectively detect gene-gene interactions associated with macular degeneration and cataract (AREDS).

关 键 词：交互效应 函数数据分析 COX比例风险模型 SNP 

分 类 号：R73[医药卫生—肿瘤]