Study On The Influence Of Population Stratification On Gene-environment Interaction Tes

Genome-wide association analysis studies have shown that genetic variants in many complex diseases are strongly associated with gene-environment interactions.It is of great theoretical and practical importance to explore the tests of geneenvironment interactions.At the same time,population stratification can have an impact on genome-wide association analysis by biasing results and presenting false positives.Therefore,exploring the effects of population stratification on geneenvironment interaction tests can help to better explore the relationship between complex diseases and G-E interaction tests.To date,fewer studies have combined the two,and the main challenges are group stratification and correlation within family data.In this paper,we propose a semi-parametric approach based on family data(PG-GE)to investigate the effect of population stratification on the geneenvironment interaction test.We use principal component analysis(PCA)to correct for population stratification and generalized estimating equations(GEE)to deal with correlations within family structure and thus establish test statistics to comprehensively test for gene-environment interactions in single variant loci.Geneenvironment interactions.We evaluated the robustness and efficiency of the PG-GE method through extensive simulations,which showed that our proposed PG-GE method could control the effect of population stratification on the G-E interaction test compared with other methods,while it could control the Type I error rate very consistently under different settings,demonstrating higher efficacy.In addition,we also performed real data analysis using the GAW17 dataset,and the results showed that the PG-GE method detected the interaction of genetic loci of ARNT and KDR genes with smoking environment,which was the most efficient and advantageous among all methods,indicating the applicability of the PG-GE method.In this study,the method was implemented by R and MATLAB software.