Current and Future Perspectives in DNA Methylation – Integrating Multi-omics Data

Abstract

Epigenome-Wide Association Study (EWAS) of human complex traits is a rapidly emerging area of research. Its purpose has been to capture the epigenetic profiles (DNA methylation) of individuals, and understanding how their epigenomes are modified throughout lifetime as well as methylation patterns associated with phenotypes and response to environment. The most widely used platform for EWAS is array-based assay, which is considered a good compromise between genome coverage and cost expense. However, this method requires more effort in quality control (e.g. identify and adjust for batch effects) and normalization in the analytical pipeline. Advances have also been made in the selection of covariates that form the best-fit model incorporating family structures (e.g. monozygotic twins study design) using the linear mixed effect model. Furthermore, within an individual, there are cross-tissue (e.g. blood, adipose, skin) differences to consider, such as tissue-shared and tissue-specific methylation changes. Together these associations are increasingly being co-analyzed with -omics data (such as gene expression and metabolomics) under statistical models (e.g. causal inference test) to give insight into novel molecular processes and regulations of disease. ????? In this talk, I will present the statistical methodology of EWAS and integrated -omics data deriving from array, RNA-seq, and metabolomics, by way of examples from my publications, which include: quality control, power estimation, methylation age, and environmental influence to the methylome (e.g. smoking behavior).