Omics field

Epigenomics

Epigenomics is the genome-wide study of chemical marks and structural states layered on DNA and chromatin — DNA methylation, histone modifications, chromatin accessibility and three-dimensional contacts — that shape which parts of an identical genome are usable in a given cell type or condition.

Questions this field can answer

Which regions of the genome are open and available to regulatory proteins in this cell type?
Which promoters and enhancers carry active or repressive histone marks?
Which CpG sites or regions differ in methylation between conditions?
Which distal regulatory element contacts which promoter?
Does a non-coding variant fall inside a regulatory element active in the relevant tissue?
How does an exposure, age or treatment associate with regulatory state?

Method

Research workflow

Established practice, from framing a question to depositing data others can reuse.

01
Study design
Fix the cell type or tissue first — epigenomes are cell-type specific, so a mixed tissue gives a mixture. Decide sample size from expected effect size, and balance case and control across processing batches and array chips or sequencing lanes.
02
Sampling
Record collection time, cell input number, fixation or freezing method, and any sorting step. Cell composition of a bulk sample is itself a biological variable that must be measured or estimated.
03
Assay and data generation
Choose the assay that matches the question: bisulfite or array methylation, ChIP-seq or CUT&RUN for histone marks, ATAC-seq for accessibility, Hi-C for contacts. Include the required inputs and controls for that assay.
04
Quality control
Check bisulfite conversion rate, control probe performance and detection p-values on arrays; for sequencing, check duplication, mitochondrial and TSS enrichment, fragment-size periodicity, FRiP and library complexity.
05
Analysis
Align to a stated assembly, normalize, then estimate and adjust for cell-type composition and batch. Call peaks or differentially methylated positions and regions, and control multiple testing genome-wide.
06
Interpretation
Association is not mechanism and not direction. A mark can follow transcription rather than cause it, and a differentially methylated region may reflect a shift in cell proportions, not regulation within a cell.
07
Reproducibility
Report assembly, blacklist, peak caller and version, normalization, covariates and the exact composition-adjustment method. Share code and processing parameters.
08
Deposition
Deposit raw and processed data in GEO or ENA with full sample metadata, and cite the accession. Follow consortium data-use terms for any restricted human dataset.

Practice

Samples, technologies and what can go wrong

Method choice sets the ceiling on what an analysis can show. Limitations are part of the method, not an afterthought.

Sample types

Sorted primary cells, the cleanest option for regulatory state
Whole blood and peripheral blood mononuclear cells, with cell-composition estimation
Fresh-frozen solid tissue and nuclei preparations
FFPE tissue, where fragmentation limits assay choice
Cell lines, reproducible but drifting with passage
Cell-free DNA, where methylation carries tissue-of-origin signal

Major technologies

Whole-genome bisulfite sequencing (WGBS): Single-base methylation across the genome; comprehensive, expensive, and bisulfite alone cannot distinguish 5mC from 5hmC.
Methylation arrays (EPIC/450K): Cost-effective measurement at a fixed set of CpG sites; excellent for cohorts, but limited to designed probes and subject to probe cross-hybridisation.
Enzymatic methyl-seq (EM-seq): Avoids bisulfite-induced degradation, giving more even coverage from less input.
ChIP-seq: Maps a specific histone modification or bound protein; needs a matched input control and a validated antibody.
CUT&RUN / CUT&Tag: Targeted cleavage approaches with much lower cell input and background than ChIP-seq.
ATAC-seq: Maps open chromatin from few cells; sensitive to mitochondrial contamination and transposase-to-cell ratio.
Hi-C and derivatives: Measures three-dimensional contacts, compartments, domains and loops at a resolution set by sequencing depth.
Long-read native methylation: Nanopore and PacBio call modifications directly on native DNA, phased onto haplotypes without conversion chemistry.

Limitations

Bulk assays report a population average; a difference can come from composition rather than regulation.
Epigenomes are cell-type and context specific, so results rarely transfer across tissues.
Cross-sectional associations cannot establish direction of causality.
Standard bisulfite methods conflate 5mC and 5hmC.
Peak calls depend heavily on antibody quality, control choice, depth and caller settings.
Array probes overlapping common variants can produce artefactual methylation differences.

Common confounders

Cell-type composition — the dominant confounder in blood and most tissues.
Age, sex and genetic ancestry, all of which shape methylation strongly.
Batch, chip and position-on-chip effects in array studies.
Smoking, medication and other exposures with large, well-documented signatures.
Post-mortem interval, ischaemic time and storage duration.
Differing input amounts and sequencing depth between groups.

Live data

Search the public record

Read-only searches against public databases. Madomic presents and explains the records; the databases named below remain their source and owner.

Madomic Research Explorer

Search public ENCODE epigenomics experiments

Live, read-only search of the ENCODE portal. Results come directly from ENCODE and are capped at eight experiments per search; open each accession to see the full experiment, its files and its quality metrics.

Searches are capped and rate limited. Nothing you type is stored. Check ENCODE data-use terms before reuse.

Reference

Glossary

Essential terms, in plain English.

CpG island: A CG-dense region, often at a promoter, where methylation typically associates with silencing.
DMP / DMR: Differentially methylated position or region between compared groups.
Beta value / M value: Methylation proportion at a site (beta, 0–1) and its logit transform (M), preferred for statistical testing.
5mC / 5hmC: 5-methylcytosine and its oxidised derivative 5-hydroxymethylcytosine, indistinguishable by ordinary bisulfite treatment.
Histone mark: A chemical modification of a histone tail, such as H3K4me3 at active promoters or H3K27ac at active enhancers.
Chromatin accessibility: How reachable DNA is to proteins and enzymes; the quantity measured by ATAC-seq and DNase-seq.
FRiP: Fraction of reads in peaks — a standard signal-to-noise quality metric.
TSS enrichment: Signal concentration around transcription start sites, used as an ATAC-seq quality check.
Blacklist region: Genomic regions with artefactually high signal that are excluded before analysis.
TAD: Topologically associating domain — a self-interacting chromatin region seen in contact maps.
Imprinting: Parent-of-origin-specific expression maintained by allele-specific methylation.
Epigenetic clock: A model estimating biological age from methylation values; a statistical predictor, not a diagnosis.

For students

Learning path and a practical activity

Everything below uses public data only. No samples, credentials or paid services are needed.

Learning path

01Learn what chromatin is and why identical genomes give different cell types.
02Learn DNA methylation chemistry and what bisulfite conversion does and does not resolve.
03Learn the meaning of the core histone marks: H3K4me3, H3K4me1, H3K27ac, H3K27me3, H3K9me3.
04Compare ChIP-seq, CUT&RUN, ATAC-seq and Hi-C by what each actually measures.
05Work through a methylation-array QC report and understand beta versus M values.
06Learn cell-type deconvolution and why it is mandatory in blood studies.
07Browse a locus in the WashU or UCSC browser with several ENCODE tracks loaded.
08Practise interpreting one non-coding variant against tissue-matched regulatory tracks.

Activity — compare regulatory state at one locus across two tissues

01Pick a gene of interest, for example GATA1, MYOD1 or ALB, and note its GRCh38 promoter coordinates from Ensembl.
02In the explorer below, search the H3K27ac mark together with a tissue term, for example “H3K27ac liver”, and open one released experiment on the ENCODE portal.
03Record the accession, assay, biosample, target and release date, then repeat with a second, clearly different tissue.
04Open the UCSC or WashU genome browser at your gene and load the matching ENCODE signal tracks for both tissues.
05Note whether the promoter and any nearby candidate enhancers show active signal in one tissue, both, or neither.
06Add an ATAC-seq or DNase track for the same tissues and check whether accessibility agrees with the histone signal.
07Write a short paragraph stating what you observed, the assembly and accessions you used, and at least two reasons the comparison could mislead — different labs, depths, antibodies, or cell composition of the tissue.

Attribution

Research resources

Authoritative public resources. Omicser is independent of each of them and links to the official source.

ENCODE Project portal

Maintained by ENCODE Data Coordination Center (Stanford University), funded by NHGRI

Uniformly processed chromatin accessibility, histone ChIP-seq, methylation and 3D genome experiments across human and mouse biosamples, with full quality metrics.

ENCODE REST API documentation

Maintained by ENCODE Data Coordination Center

The official programmatic search interface used by the explorer on this page.

Roadmap Epigenomics Project

Maintained by NIH Roadmap Epigenomics Consortium

Reference epigenomes and chromatin-state segmentations for primary human tissues and cells.

NCBI GEO

Maintained by US National Center for Biotechnology Information

Primary archive for submitted methylation array, ChIP-seq and ATAC-seq datasets with their sample metadata.

EMBL-EBI ArrayExpress / BioStudies

Maintained by EMBL-EBI

European archive of functional genomics studies, including epigenomic experiments and their protocols.

EWAS Catalog

Maintained by MRC Integrative Epidemiology Unit, University of Bristol

Curated published epigenome-wide association study results, useful for checking whether a CpG has been reported before.

WashU Epigenome Browser

Maintained by Wang Lab, Washington University in St. Louis

Visualising many epigenomic tracks and long-range interactions at a chosen locus.

UCSC Genome Browser

Maintained by UCSC Genomics Institute

Loading regulatory, conservation and variant tracks alongside your region of interest.

Research and education use only. This page is not a clinical diagnosis, medical advice or a laboratory result. Verify every finding in the originating database and in the peer-reviewed literature before drawing conclusions.