Poor Statistical Power in Population-based Association Study of Gene Interaction

Abstract Background Statistical epistasis, a significant “gene interaction” in genetic association studies, means the nonadditive effects between the polymorphic sites on two different genes affecting the same phenotype. In genetic association analysis of complex traits, nevertheless, the researchers haven’t found statistical epistasis so far.Methods To explore the fundamental reason for this question, this study developed a modeling and computational method focusing on the statistical power and then calculated and compared the statistical power under different hypothesis testing scenarios.Results The results showed that under the parameters of large-scale genetic association analysis like genome-wide association studies, the power of the hypothesis test for detecting the statistical epistasis between different genetic marker genes was close to zero. From the results of this investigation, we can conclude that the lack of statistical epistasis in genetic association studies is an inevitable result owing to the statistical principles of the genetic association studies and the inherent characteristics of the research methods, which cannot be changed by the innovation of technical means.Conclusions It is time-consuming and hopeless to seek for the statistical epistasis in genetic association analysis of complex traits without decomposing complex traits of multiple genes into intermediate traits.