Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 430 - Challenges and Recent Advances in High-Dimensional Mediation Analysis
Type: Invited
Date/Time: Thursday, August 6, 2020 : 10:00 AM to 11:50 PM
Sponsor: Section on Statistics in Epidemiology
Abstract #308046
Title: Variance Component Tests of Multivariate Mediation Effects Under Composite Null Hypotheses
Author(s): Yen-Tsung Huang*
Companies: Academia Sinica
Keywords: composite null hypothesis; intersection-union test; joint significance test; mediation analyses; normal product distribution; variance component test

Mediation effects are determined by two associations: one between an exposure and mediators and the other between the mediators and an outcome conditional on the exposure. The test for mediation effects is conducted under a composite null hypothesis, i.e., either one of the two associations is zero or both are zeros. Without accounting for the composite null, the Type I error rate within a study containing a large number of multi-mediator tests may be much less than the expected. We propose a novel test to address the issue. For each mediation test, we examine the associations using two separate variance component tests. Assuming a zero-mean working distribution with a common variance for the element-wise associations, score tests for the variance components are constructed. We transform the test statistics into normally distributed statistics under the null and account for the composite null by adjusting for the variances of the statistics. Advantages of the proposed test are illustrated in simulation studies and a data application where we analyze lung cancer data from TCGA to investigate the smoking effect on gene expression through DNA methylation in 15,114 genes.

Authors who are presenting talks have a * after their name.

Back to the full JSM 2020 program