To address the issue of replicability of findings, future studies will need to be more standardized to maximize the chance of replication, including best practices for handling data. Relevant study design issues include batch effects, normalization of signals, analysis of confounding variables, ascertainment bias, sample processing, genotype by methylation confounding (GxMeth) and controlling for cell mixture. Fortunately good techniques exist to address these. The inclusion of other biochemical markers of exposure such as urine or serum cotinine is also recommended.
