Influence Analysis Of Weight Selection Of Aggregated Cauchy Association Test In Rare Variation Association Test

Existing studies have shown that rare variants are closely related to complex diseases or traits,so it is of great significance to develop multi-locus association analysis methods for rare variants for the study of complex traits or disease mechanisms.Aiming at the problems of limited application scope and low power of existing multi-locus association analysis methods,this paper developed a multi-locus statistical analysis method for rare variants with better applicability and higher power,and specifically studied the following contents:Firstly,based on the original Beta weighting method,this paper explored the aggregated Cauchy association test(ACAT)weighting method from four aspects,including P value weighting method based on external data,allele frequency and annotation principal component combined weighting method,improved SKAT weighting method and WST weighting method.The theoretical distribution properties of these test statistics and the method of P values calculation are given.Secondly,the effectiveness of the method is demonstrated through the type I error rate of the selected method and the comparison of its power with the original weighting method.The results show that,from the test of the type I error rate,when the significance level is 0.01 and0.05,the type I error rate of the proposed method are controlled,that is,they are all within the corresponding confidence interval.The effects of sample correlation,causal variation,effect size,effect direction and minor allele count(MAC)on the power of each weighting method were investigated.The results showed that the power gradually increased with the increase of sample size.Finally,the proposed method was applied to the real GWAS data of P53 gene and environmental stress and nerve sensitivity clusters to further verify the effectiveness and robustness of the proposed method.