A0638
Title: Mediation analysis with high dimensional exposures and confounders
Authors: Qi Zhang - University of New Hampshire (United States) [presenting]
Abstract: High-dimensional mediation analysis has been receiving increasing popularity, largely motivated by the scientific problems in genomics and biomedical imaging. Previous literature has primarily focused on mediator selection for high-dimensional mediators. The aim is to estimate and infer the overall indirect effect of high dimensional exposures and high dimensional mediators. MedDiC, a novel debiased estimator of the high dimensional overall indirect effect, is proposed based on the difference-in-coefficients approach. The proposed method is evaluated using intensive simulations, and it is found that MedDiC provides valid inference and offers higher power and shorter computing time than the competitors for both low-dimensional and high-dimensional exposures. MedDiC is also applied to a mouse f2 dataset for diabetes study and a dataset composed of diverse maize inbred lines for flowering time, and MedDiCyields is shown to provide more biologically meaningful gene lists, and the results are reproducible across analyses using different measures of identical biological signal or related phenotype as the outcome.
