Epigenetic analysis on Illumina EPIC arrays

Go Back


Epigenetics is one of the hottest fields in life sciences. From academic publications to the mainstream media, it has the ability to fill in some of the gaps in scientific knowledge and has huge potential in human medicine.

What is epigenetics?

Derived from the Greek language and coined by Waddington in 1942, it can be described as the ‘study of the change in gene expression without altering the DNA sequence’. In other words; a change in the phenotype without affecting the genotype.  At least three mechanisms have been identified that stand out as the main drivers of such a change: histone modification, non-coding RNA (ncRNA)-associated gene silencing and DNA methylation. It is the latter that we at Oxford Genomics Centre have taken a keen interest in studying over the past few years.

DNA methylation plays an important and dynamic role in regulating gene expression. It allows cells to acquire and maintain a specialised state, suppresses the expression of viral and non-host DNA elements, and facilitates response to environmental stimuli. Aberrant DNA methylation (hyper- or hypomethylation) and its impact on gene expression have been implicated in many disease processes, including cancer, neurological disorders, metabolic disease, aging, and development (1, 2, 3).

DNA Methylation refers to the addition of a methyl (CH3) group to the 5th carbon on a cytosine ring giving rise to 5-methylcytosine (5mC). The whole process is mediated by DNA methyltransferases (DNMTs) (Fig 1).

In a scenario where cytosine is followed by a guanine in a liner 5’ to 3’ direction such dinucleotide is commonly referred to as CpG island. Human genome contains ~1% of these sites, whilst about 75% of them have 5-methylcytosine. This is very significant as CpG islands have been identified in or near a large number of promoters. However, CpG sites are also found outside of CpG islands (Fig 2). Should a large number of CpG sites within a promoter be methylated, the result is gene silencing. This has large implication in many diseases and cancer development (4).

One traditional method of studying methylation is by using bisulfite treatment. However new discoveries have shown the limitation of such an approach to discriminate between 5mC and 5hmC; the first intermediary in the demethylation of cytosine (Fig 3). The new TrueMethyl oxidative bisulfite kits from CEGX overcomes the shortfall of the traditional bisulfite analysis by oxidising 5hmC to 5fC which is then converted to uracil using bisulfite. The identity of 5hmC from 5mC can be discriminated by comparison to a traditional bisulfite protocol (Fig 4 a & b).

Image gallery

Anew study suggests that 5hmC is an active epigenetic marker stably maintained after the global reprogramming event during early embryonic development (5). Furthermore, in a recent nature paper, researchers at the Geisel School of Medicine, USA, demonstrated the use of oxidative bisulfite (oxBS) treatment to assess 5hmC in glioblastoma samples (6). Interestingly, the level of 5hmC could be correlated with patient survival, clearly demonstrating the potential utility of 5hmC as a prognostic biomarker (this article was subsequently mentioned by Director of the NIH himself, Francis Collins, posting on his blog).

For high-throughput investigation, the Illumina Infinium MethylationEPIC BeadChip is definitely the platform of choice (Fig 5). Designed as an 8 sample bead array, kits are available for 16, 32 and 96 samples, allowing an investigator to interrogate over 850,000 methylation sites quantitatively across the genome at single-nucleotide resolution (> 90% of the original CpGs from Human Methylation 450BeadChip plus an additional 350,000 CpGs in enhancer regions). Multiple samples can be analysed in parallel to deliver high-throughput power whilst minimising the cost per sample.

Infinium HD array technology enables content selection independent of bias-associated limitations often associated with methylated DNA capture methods. The result is a pan-enhancer and coding region view of the methylome that can be used for epigenome-wide association studies on a variety of human tissues.

Through small modification of the current methylation protocol it was demonstrated that even the FFPE samples can be processed to produce robust and reliable results (Table 1).

Table 1.  Comparative Infinium Methylation Data Quality Metrics-Standard vs.  FFPE

Methylation BeadChip Standard Protocol FFPE Protocol

(Technical Replicates)

r2 > 98% r2 > 98%
Number of sites detected* > 96% > 90%

* Based on non-cancer samples, recommended sample input amounts of high-quality DNA as confirmed by Pico Green and following all other Illumina recommendations as per User Guides.

For small scale projects the most common method of data analysis is with Illumina GenomeStudio Methylation Module, enabling visualisation of heatmaps, chromosomal coordinates, percent GC, location in a CpG Island, and methylation beta values, whereas large scale studies mainly utilise packages that work in the software framework of R (7, 8).

In addition, data analysis workflows are now available through the DNAnexus cloud-based genome informatics and data management platform, offering TrueMethyl users another alternative for analysing and visualising oxBS data.

Authors: Bobby Bojovic, Christine Blancher



  1. Portela A, Esteller M. Epigenetic modifications and human disease. Nat Biotech. 2010; 28:1057-1068.
  2. Rakyan VK, et al. Epigenome-wide association studies for common human diseases. Nat Rev Genetics. 2011; 12(8):529-541.
  3. Wahl S, et al, Epigenome-wide association study of body mass index, and adverse outcomes of adiposity. Nature. 2017 (541): 81-86.
  4. Mendizabal I, et al. Body-hypomethylated human genes harbor extensive intragenic transcriptional activity and are prone to cancer-associated dysregulation. NAR. 2017 Jan, 23 [Epub ahead of print]
  5. Lin IH, et al. Correlated 5-Hydroxymethylcytosine (5hmC) and Gene Expression Profiles Underpin Gene and Organ-Specific Epigenetic Regulation in Adult Mouse Brain and Liver. PLoS One. 2017 Jan 26;12(1)
  6. Johnson KC, et al. 5-Hydroxymethylcytosine localizes to enhancer elements and is associated with survival in glioblastoma patients. Nat Commun. 2016 Nov 25;7:13177.
  7. Assenov Y, et al. Comprehensive analysis of DNA methylation with RnBeads. Nat Methods. 2015;11(11):1138-1140.
  8. Morris TJ, et al. ChAMP: 450K chip analysis methylation pipeline. Bioinformatics. 2014;30(3):428-430.