Event Description
The Department of Epidemiology and Biostatistics welcomes Jincheng Shen, Postdoctoral Fellow, Department of Biostatistics at Harvard University, who will present:
Hypothesis Testing for Mediation Effects in the Context of Epigenomic Studies.
This talk will aim to study the underlying mechanisms of whether the effect of environmental exposures on disease outcome is mediated through genomic and epigenomic markers, e.g., DNA methylations. Mediation analysis provides a powerful tool for mechanism studies, and has gained great success in social science research. There has been growing interests in employing mediation analysis technologies in genetic and epigenetic studies to look at causal pathways and mechanisms. We have worked on addressing challenges of developing proper tests for single mediator, as well as multiple mediators in epigenomic setting. The proposed methods can effectively account for the composite nature of the null hypothesis and allow for weak and sparse signals. Extensive simulation studies have been conducted to assess the type I error rates and powers under various practical settings. The proposed tests have also been applied to the Normative Aging Study and identified putative DNA methylation CpG sites as mediators in the causal pathway of smoking behavior to lung functions. |