Abstract:<span>In genome-wide epigenetic studies, it is of great scientific interest to assess whether the effect of an exposure on a clinical outcome is mediated through DNA methylations. However, statistical inference for causal mediation effects is challenged by the fact that one needs to test a large number of composite null hypotheses across the whole epigenome. Two popular tests, the Wald-type Sobel's test and the joint significant test using the traditional null distribution are underpowered and thus can miss important scientific discoveries. In this article, we show that the null distribution of Sobel's test is not the standard normal distribution and the null distribution of the joint significant test is not uniform under the composite null of no mediation effect, especially in finite samples and under the singular point null case that the exposure has no effect on the mediator and the mediator has no effect on the outcome. Our results explain why these two tests are underpowered, and more importantly motivate us to develop a more powerful divide-aggregate composite-null test (DACT) for the composite null hypothesis of no mediation effect by leveraging epigenome-wide data. We adopted Efron's empirical null framework for assessing statistical significance of the DACT test. We showed analytically that the proposed DACT method had improved power, and could well control Type I error rate. Our extensive simulation studies showed that, in finite samples, the DACT method properly controlled the Type I error rate and outperformed Sobel's test and the joint significance test for detecting mediation effects. We applied the DACT method to the U.S. Department of Veterans Affairs Normative Aging Study, an ongoing prospective cohort study which included men who were aged 21 to 80 years at entry. We identified multiple DNA methylation CpG sites that might mediate the effect of smoking on lung function with effect sizes ranging from –0.18 to –0.79 and false discovery rate controlled at the level 0.05, including the CpG sites in the genes AHRR and F2RL3. Our sensitivity analysis found small residual correlations (less than 0.01) of the error terms between the outcome and mediator regressions, suggesting that our results are robust to unmeasured confounding factors. Supplementary materials for this article, including a standardized description of the materials available for reproducing the work, are available as an online supplement.</span>

Powerful Large-scale Inference in High Dimensional Mediation Analysis

An Efficient Testing Procedure for High-Dimensional Mediators with FDR Control

Large-Scale Hypothesis Testing for Causal Mediation Effects with Applications in Genome-wide Epigenetic Studies

Testing Mediation Effects in High-Dimensional Microbiome Data with False Discovery Rate Control

AMDP: an Adaptive Detection Procedure for False Discovery Rate Control in High-Dimensional Mediation Analysis.

Estimating and Testing High-Dimensional Mediation Effects in Epigenetic Studies

A Dual-Penalized Approach to Hypothesis Testing in High-Dimensional Linear Mediation Models

DP2LM: leveraging deep learning approach for estimation and hypothesis testing on mediation effects with high-dimensional mediators and complex confounders

Mediation analysis in longitudinal study with high-dimensional methylation mediators

Estimation and inference for the indirect effect in high-dimensional linear mediation models

Individualized Dynamic Mediation Analysis Using Latent Factor Models

Power of Testing for Exposure Effects under Incomplete Mediation

High-dimensional confounding in causal mediation: a comparison study of double machine learning and regularized partial correlation network

Controlling false discovery rate for mediator selection in high-dimensional data

Exploratory Mediation Analysis with Many Potential Mediators

Non-linear Mediation Analysis with High-dimensional Mediators whose Causal Structure is Unknown

A framework for mediation analysis with massive data

High dimensional mediation analysis with latent variables

Testing High-Dimensional Mediation Effect with Arbitrary Exposure-Mediator Coefficients

Mediation Analysis with Mendelian Randomization and Efficient Multiple GWAS Integration

Dissecting the colocalized GWAS and eQTLs with mediation analysis for high-dimensional exposures and confounders