A Statistical Model Based On Gene-environment Interaction In Type 2 Diabetes Mellitus

Objective With the prevalence of diabetes worldwide,diabetes has become one of the global public health concerns.Diabetes,a multifactorial disease,is closely related to environmental factors and genetic factors.In the genetic mechanism,multi-factor disease is controlled by multiple genes.Multifactorial disease is the result of complex interactions between multiple genes or gene-environment interactions.Exploring the influence of interaction on diseases has become one of the important contents of genetic epidemiology.So far,GWAS research on many diseases have been carried out great achievements.However,the results fell far short of expectations.SNPs for specific diseases account for only a small fraction of genetic variation.There are several reasons for this kind of missing heritability.One of the main reasons may be neglect of gene-environment interactions.Therefore,detection of gene-environment interaction has an important theoretical and practical significance for improving the genetic interpretation of multi-factor diseases,building reasonable genetic risk assessment models,developing personalized drug targets for diagnosis and treatment,etc.Methods Based on the case-control trial design and the principle of linkage disequilibrium,this study intends to construct a new statistical inference method of gene-environment interaction to conduce to gene location of diabetes.The effectiveness of the new method is demonstrated through the theoretical derivation and computational simulation.The statistical characteristics of the new method can be directly verified by computational simulation.Compared with several common methods,the practicability and effectiveness are further demonstrated.Results A new statistic for detecting gene-environment interaction was obtained,approximately following distribution.Statistical simulation and case verification show that ?²: ?The probability of the constructed statistic T_IH making type ? errors is stable around the given test level.?With a certain sample size,the test efficiency of the constructed statistic T_IH increases with the increase of interaction effect.The test efficiency of the constructed statistic A increases with the increase of the sample size under a given interaction effect size.?Under the same conditions,the test efficiency of the constructed statistic T_IH is significantly higher than that of bifurcation analysis and logistic regression model.?The constructed statistic T_IH also fully shows its superiority in the example verification.Conclusion Through the comparison of simulation,the test efficiency of the new method is better than that of the commonly used bifurcation analysis and logistic regression model.The new method for detecting gene-environment interaction has good statistical performance,and can be applied in practice.