病例-对照设计的似然比罕见变异关联性检验的构建和模拟评价

发布时间：2018-02-11 22:58

本文关键词： 罕见变异关联性分析 logistic混合模型似然比检验　出处：《郑州大学学报(医学版)》2017年02期 　论文类型：期刊论文

【摘要】：目的:发展适用于病例-对照设计的似然比罕见变异关联性分析方法。方法:在logistic混合模型的框架下基于PQL算法建立伪数据,将二分类表型转化为连续型表型的关联性分析,然后借助线性混合模型的方差成分检验执行关联性分析。采用Monte Carlo模拟评价该方法的有效性,并与现有方法进行对比。结果:模拟显示,在不同情况下包括似然比检验在内的所有统计检验都能有效控制Ⅰ型错误;在效应方向相同情况下,Burden检验、SKAT-O和Mi ST的统计效能高;在效应方向不同的情况下似然比检验优于其他方法。结论:基于logistic混合模型和PQL算法的似然比检验可有效用于病例-对照设计的罕见变异关联性分析。
[Abstract]:Objective: to develop a method of likelihood ratio rare variant correlation analysis for case-control design. Methods: pseudo data based on PQL algorithm was established under the framework of logistic mixed model to transform the two phenotypes into continuous phenotypic correlation analysis. Then the variance component test of the linear mixed model is used to perform the correlation analysis. The effectiveness of the method is evaluated by Monte Carlo simulation and compared with the existing methods. All statistical tests, including likelihood ratio test, can effectively control type I errors under different conditions, and have high statistical efficacy of SKAT-O and Mi St under the same effect direction. Conclusion: the likelihood ratio test based on logistic mixed model and PQL algorithm can be effectively used in the rare mutation correlation analysis of case-control design.
【作者单位】：徐州医科大学公共卫生学院流行病与卫生统计学教研室;南京医科大学公共卫生学院生物统计学系;
【基金】：国家自然科学基金项目81402765 国家统计局全国统计科学研究项目2014LY112
【分类号】：R394

【相似文献】