| Summary: | 博士 === 國防醫學院 === 生命科學研究所 === 104 === Conventional genome-wide association studies (GWAS) have successfully identified many important genetic variants in few complex human traits. However, there is still a large heritability gap in most human phenotypes. The ‘missing heritability’ has been suggested to be due to lack of technologies detected gene–gene and gene–environment interactions. An individual trial has often had insufficient sample size, and meta-analysis is a common method for increasing statistical power. However, sufficient detailed individual data is difficult to obtain. Here, this study provides two methods to detect gene–gene and gene–environment interactions using summary data, respectively. The first method is based on traditional meta-regression, and this study improves its performance; the second method is a novel Markov chain Monte Carlo-based method, called ‘Epistasis Test in Meta-Analysis’ (ETMA). This study defined a series of conditions to generate simulation data and tested the type I error rates in these two methods, and these two methods both yielded acceptable type I error rates. Moreover, this study applied these two methods to few real meta-analysis data sets. Improved meta-regression successfully found a gender dependent effect of angiotensin-converting enzyme insertion/deletion on chronic kidney disease; ETMA successfully found significant gene–gene interactions in the polycyclic aromatic hydrocarbon metabolism pathway and the renin–angiotensin system, with strong supporting evidence. Finally, an R package, etma, for the detection of epistasis in meta-analysis has been developed and published in Comprehensive R Archive Network (https://cran.r-project.org/web/packages/etma).
|
|---|
