Rodger and Chatterjee Clinical Epigenetics (2017) 9:5 DOI 10.1186/s13148-017-0313-y SHORT REPORT Open Access The epigenomic basis of common diseases Euan J. Rodger1,2* and Aniruddha Chatterjee1,2* Abstract A report of the 6th Epigenomics of Common Diseases Conference held at the Wellcome Genome Campus in Hinxton, Cambridge, UK, on 1–4 November 2016. Introduction to investigate a large number of CpGs across large num- Epigenetic modification provides a stable mechanism by bers of patients and controls to detect aberrant methyla- which cells with the same genotype can modulate their tion signals at a population level [1]. Further, EWASs are gene expression and exhibit different phenotypes. In the an ideal platform to tap into large international resources past two decades, excellent progress has been made to and compare multiple datasets with custom-generated profile these modifications and our understanding of EWAS data. Examples of well-curated epigenomic data- epigenetic marks has surpassed beyond the basic sets include the International Human Epigenome Consor- phenomenon of cellular heterogeneity. It is now estab- tium (IHEC), the EU-funded BLUEPRINT project, and lished that epigenetic marks are altered in almost all the International Cancer Genome Consortium (ICGC). common human diseases. The Epigenomics of Common Although EWASs have been in use for several years now Diseases meeting, 1–4 November 2016, provided an and thousands of datasets and several analytical tools have account of the progress made in this area and also indi- been reported, there is still a need to understand the po- cated future areas that are yet to be addressed. Although tential biases and the nature of factors that could influence disease focussed, several other aspects were discussed the interpretation of results. More sophisticated tools need that are relevant to epigenetics as a field, including cellu- to be developed to account for these factors. One observa- lar heterogeneity, epigenomic association studies, emer- tion to come out of this meeting, based on the commen- ging concepts in cancer epigenetics and new innovative tary of multiple speakers, was that EWASs require robust techniques of broad application (such as single-cell ana- analytical tools to detect epigenetic variants of interest lysis and epigenomic editing approaches). Here, we pro- and to adjust for confounding factors such as genetic vide a brief report of some of the key ideas and themes effects and cellular heterogeneity. discussed in this meeting and based on these, we specu- Several speakers presented vignettes of many interest- late on future research directions. ing EWAS findings, including Stephan Beck (University College London, UK) who shared some “new twists” in EWAS analytics and started off with a plea for authors A needle in the haystack: insights from epigenome-wide and editors alike to ensure EWAS papers include CpG association studies numbers. This oversight, he suggested, was akin to pub- Recent advances in high-throughput DNA analysis now lishing a genome-wide association study without includ- enable researchers to examine epigenetic modifications ing rs numbers for single-nucleotide polymorphisms. across the genome, primarily DNA methylation marks, Methylation of a CpG site and its relationship to the for association with numerous disease phenotypes. As expression of a corresponding or nearby gene is very such, epigenome-wide association studies (EWASs) have context specific. For example, the methylation of a par- been fruitful in their findings but they also harbour their ticular gene promoter is likely to be very different to the own unique challenges. EWASs provide an opportunity methylation of the gene body. Methylation in either of * Correspondence: [email protected]; these elements could have a different function for the [email protected] specificity of the corresponding transcription [2]. There- 1Department of Pathology, Dunedin School of Medicine, University of Otago, fore, providing the methylation of a gene as a whole is Hanover Street, P.O. Box 56, Dunedin 9054, New Zealand Full list of author information is available at the end of the article almost meaningless if the context (i.e. the CpG number) © The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Rodger and Chatterjee Clinical Epigenetics (2017) 9:5 Page 2 of 6 is not described. In their analysis of >700 haemopoietic known risk factors. The genes associated with three of these effector cell methylomes from monozygotic twins dis- loci (TXNIP, ABCG1,andSREBF1) have biological plausi- cordant for type 1 diabetes, Beck and colleagues were bility in the context of type 2 diabetes. In an EWAS investi- only able to identify a single significant differentially gating adiposity, methylation markers at 187 loci were methylated CpG position (DMP). Undeterred, they found to be associated with body mass index (BMI) [10]. shifted their focus from differences in mean methylation These loci were enriched for functional regulatory elements and used a novel approach to detect differentially vari- and gene promoters that mapped to biologically plausible able positions (DVPs). The diabetes-associated DVPs pathways. However, Mendelian randomisation to test for were temporally stable and mapped onto regulatory cir- causal relationships indicated that the changes in methyla- cuits involved in cell cycle and immune cell metabolism. tion were more likely to be a consequence rather than a Similarly, they also applied this approach to methylomes cause of BMI. from monozygotic twins discordant for rheumatoid arth- ritis and also to the discovery of DNA methylation and Epigenetics in cancer gene expression variability in normal blood cells. Sub- The mechanisms involved in the hallmark properties of stantial DNA methylation and gene expression variation a cancer cell are well described and continue to be heav- in normal blood cells has also been reported in recent ily investigated [11]. Until recently, cancer was often per- genome-wide studies [3–5]. Therefore, it is important to ceived as only a disease of the genome. However, it is consider the normally occurring variation in methylomes now established that epigenetic changes are present in for detecting differential methylation signals. Beck also all human tumours and the fact that cancer is a disease presented several bioinformatic tools for EWAS inter- of both the genome and epigenome is gaining recogni- pretation such as eFORGE (http://eforge.cs.ucl.ac.uk), tion. The analysis of thousands of cancer genomes re- which identifies tissue or cell-type specific signals from vealed that epigenomic regulator genes were often Illumina 450K methylation array data [6], EpiDISH (freely mutated in many cancers [12]. Epigenetic alterations available from https://github.com/sjczheng/EpiDISH), have been shown to cooperate with genetic alterations which can be used for epigenetic dissection of heterogen- to drive the cancer phenotype [13]. These changes can eity within a sample, and CORALINA (comprehensive be used as biomarkers of disease state and the poten- guide RNA library generation through controlled nuclease tially reversible nature of epigenetic aberrations has been activity), a universal method for generating guide RNA an alluring prospect for the field of epigenetic therapies. libraries for large-scale CRISPR-based genomic and epige- In her talk on epigenetics of the cancer microenviron- nomic screening [7]. ment, Susan Clark (Garvan Institute of Medical Re- Bill Cookson (Imperial College London, UK) presented search, Sydney, Australia) began by quoting Stephen an EWAS of total serum immunoglobulin E (IgE), which Paget’s 1889 “soil and seed hypothesis”—“When a plant is a central mediator in asthma and atopy, in peripheral goes to seed, its seeds are carried in all directions; but blood leukocyte methylomes [8]. Of the 36 loci showing they can only live and grow if they fall on congenial an association between methylation and IgE concentration, soil”. The focus of Clark’s investigation was on pro- several loci were annotated to genes related specifically to tumorigenic cancer-associated fibroblasts (CAFs), which eosinophil function (e.g. IL5RA), which is consistent with in vitro studies have shown, were not just transiently ac- the presence of activated eosinophils in atopic subjects. tivated by signalling from tumour cells but they retain Notably, a monoclonal antibody targeted to IL5RA is their phenotype even when tumour cell stimuli were re- currently in phase 3 trials for the treatment of severe moved. Whole-genome bisulfite sequencing (WGBS) asthma. Cookson reiterated the point that cell hetero- and RNA-Seq were used on CAFs from prostate cancer geneity and genetic factors need to be accounted for and compared to matched non-malignant prostate fibro- in EWASs. The WCGNA package (freely available blasts (NPFs). This comparison showed a large number of from https://labs.genetics.ucla.edu/horvath/Coexpression