Hi-GLMM的应用对提高疾病性状基因组关联分析检测效率的影响  

Effects of Hi-GLMM application on improving the detection efficiency of genome-wide association analysis for disease traits

作　　者：徐亚楠 杨理昂 杨润清[2] 李淑玲[1] XU Yanan;YANG Li'ang;YANG Runqing;LI Shuling(School of Life Sciences,Northeast Agricultural University,Harbin 150030,China;Research Centre for Aquatic Biotechnology,Chinese Academy of Fishery Sciences,Beijing 100141,China)

机构地区：[1]东北农业大学生命科学学院,哈尔滨150030 [2]中国水产科学研究院生物技术研究中心,北京100141

出　　处：《东北农业大学学报》2024年第8期208-221,共14页Journal of Northeast Agricultural University

基　　金：国家自然科学基金面上项目(32372848)。

摘　　要：为降低计算复杂度,研究提出一种基于广义线性混合模型新型简化算法——分层广义线性混合模型(Hi-GLMM)。采用基因组最佳线性无偏预测方法(GBLUP)对正态基因组变量估计,获得基因组估计育种值(GEBVs),通过广义最小二乘方法在全基因组范围内逐个关联检验每个标记的加性效应。模拟与实际结果表明:Hi-GLMM可有效控制统计错误,提高二分类性状基因检测效率,并成功检测到与多种疾病显著相关的遗传位点。To reduce the computational complexity,this study proposed a novel simplified algorithm based on the generalized linear mixed model(GLMM)—the hierarchical generalized linear mixed model(Hi-GLMM).This method employed the genomic best linear unbiased prediction(GBLUP)approach to estimate standard genomic variables and obtain genomic estimated breeding values(GEBVs).It then tested the additive effects of each marker in the genome using the generalized least squares method in a stepwise manner.Both simulation and real-world results demonstrated that Hi-GLMM could effectively control statistical errorrs,significantly enhance the efficiency of gene detection for binary traits,and identify genetic loci associated with various diseases.

关 键 词：分层广义线性混合模型 基因组最佳线性无偏预测 基因组估计育种值 二分类性状 

分 类 号：Q348[生物学—遗传学]