Epigenome-wide association studies in the era of meta-epigenomics. J. M. Greally1, N. A. Wijetunga1, F. Delahaye2, Y. M. Zhao2, A. Golden1, J. C. Mar3, F. H. Einstein2 1) Department of Genetics, Albert Einstein College of Medicine, Bronx, NY; 2) Department of Obstetrics & Gynecology and Women's Health, Albert Einstein College of Medicine, Bronx, NY; 3) Department of Computational and Systems Biology, Albert Einstein College of Medicine, Bronx, NY.

   The epigenome-wide association study (EWAS) is increasingly employed to assess the possible role of epigenetic dysregulation in the determination of phenotypes and disease. As part of ongoing EWAS, we explored the extent of variability of the epigenome between normal individuals. We found substantial variation of DNA methylation to occur in human CD34+ hematopoietic stem and progenitor cells (HSPCs) from different neonates without apparent disease, born at term with appropriate weight for gestational age. Empirical annotation of the genome of this cell type using Roadmap in Epigenomics reference data was performed, using Segway and Self-Organising Map (SOM) analyses. Promoter, enhancer, transcribed and repressed regions were identified and annotated, revealing among other findings that so-called CpG island shores encode poised enhancers in CD34+ HSPCs.
    We then tested where DNA methylation variability was occurring relative to these empirically-derived annotations. We found the epigenetic variability to be targeted to functional sequences, candidate promoters and enhancer sequences, with a discrimination of enhancer variability proximal to genes expressed at lower levels. The enrichment of variability at loci with intermediate DNA methylation values, occurring at apparently poised enhancers (encoded by H3K4me1 and H3K27me3), suggests cell subpopulation heterogeneity between individuals as the basis for the epigenomic variability observed.
    We find that even in purified cell types the population of cells has a mixture of epigenomes present, which could be described as a meta-epigenome. We developed new meta-epigenomic approaches to study the CD34+ HSPCs. The variability of DNA methylation at candidate enhancers was low for housekeeping genes but high for genes associated with leukocyte differentiation. We also adapted an approach used for transcriptional studies to infer the number of cell subpopulations in the CD34+ HSPC pool.
    Recognition of the meta-epigenomic structure of cell pools studied by EWAS approaches establishes a foundation for more informed design and interpretation of EWAS, more rational estimates of statistical power and potentially new analytical approaches that take into account the unusual variability structure of the data, as we will describe.

You may contact the first author (during and after the meeting) at