Source code for "Controlling bias and inflation in epigenome- and transcriptome-wide association studies using the empirical null distribution"

Epigenome- and transcriptome-wide association studies are
      becoming increasingly common. Here we show that such studies are
      not only prone to significant inflation but also to bias of
      test-statistic, an unrecognized phenomenon introducing spurious
      findings if left unaddressed. Neither GWAS-based methodology or
      state-of-the-art confounder adjustment methods are able to
      completely remove bias and inflation. We propose a Bayesian
      method for the control of bias and inflation in EWAS/TWAS based
      on estimation of the empirical null distribution.  Simulation
      studies and analyses of empirical data demonstrate that our
      method maximizes power while properly controlling the false
      positive rate. Finally, we illustrate the utility of our method
      in large-scale EWAS/TWAS meta-analyses of age and smoking.