Two-stage Genetic Association Studies While Considering Genotype Uncertainty

Genome-wide association study?GWAS?is commonly used in the research of human genetic diseases.The purpose of GWAS is to find out the risk genes using hypothesis test.Since the first publication of GWAS application essay in 2005,this method has helped scientists to screen genovariations related to many complex genetic diseases such as Diabetes,Rheumatoid Arthritis and Schizophrenia.Sometimes,genotypes cannot be precisely decoded and only genotype probability is obtained.In this case,genotype imputation is usually used to determine the genotype of each individual,but inaccurate imputation might affect the results.The two-stage analysis is a common method in association analysis.Considering genotype uncertainty,this paper develops a robust test and distributions based on MAX3 method in two-stage association analysis.Using computer simulation,this paper evaluates the statistical efficiency of the proposed statistic,and applied this statistic in an illustration.Simulation result shows that the F_J^MAX statistic basis on genotype probability is more robust than?9?62?0?under REC,ADD and DOM models,and is more powerful under small minor allele frequency compare with?9?62?0?.In general,the proposed statistic is useful for genotype uncertainty in two-stage association analysis while the underlying genetic model is unknown.