What Is Epigenetics?

Epigenetics is the branch of biology that studies how gene activity is regulated by chemical modifications to DNA and to the proteins packaged with it, rather than by changes to the DNA sequence itself. The National Human Genome Research Institute (NHGRI) defines it as a field of study focused on changes in DNA that do not involve alterations to the underlying sequence. It explains why a nerve cell and a muscle cell can carry the same genome and still behave completely differently, and why environmental exposures, age and disease can leave measurable, sometimes lasting, marks on how genes are used. It is also one of the most over-interpreted terms in popular science, so this guide separates what the evidence supports from what it does not. It covers the definition, the three main molecular mechanisms (DNA methylation, histone modification and non-coding RNA), how epigenome-wide studies work, the history of the field, how epigenetics research is funded, training paths, and where it intersects with research administration.

What Is Epigenetics?

The prefix “epi-” means “on top of” or “above.” Epigenetics studies the layer of regulation that sits on top of the genome’s base sequence. Every cell in a person’s body (with a few exceptions, such as mature red blood cells) carries essentially the same DNA, but each cell type reads a different subset of its genes. Epigenetic mechanisms are a major part of how that selective reading is established, maintained and, in some cases, passed on when a cell divides.

Three terms are worth keeping distinct:

• Epigenetics is the field of study, and also a shorthand for the regulatory mechanisms themselves.
• An epigenetic mark is a specific chemical modification, such as a methyl group on a cytosine or an acetyl group on a histone.
• The epigenome is the complete set of such marks across a genome in a given cell type, tissue or condition. NHGRI notes that the complete set of epigenetic changes within a genome is called the epigenome. Unlike the genome, which is nearly identical across a person’s cells, the epigenome differs by cell type and changes over time.

The US Centers for Disease Control and Prevention describes the practical, public-health framing: how your behaviors and environment can cause changes that affect the way your genes work. The CDC’s examples include differences between nerve and muscle cells that share the same DNA, changes in DNA methylation at the AHRR gene associated with smoking, and the long-lasting methylation differences observed in people exposed prenatally to the Dutch Hunger Winter famine. The CDC also states that such changes are reversible, noting that former smokers can regain methylation levels similar to nonsmokers, in some cases in under a year.

Epigenetics sits alongside this site’s other life-science discipline guides. Genetics studies genes and inheritance, genomics studies whole genomes, and molecular biology studies the molecular machinery of the cell. Epigenomics, the genome-wide study of epigenetic marks, is one of the recognised subfields of genomics, and it relies heavily on bioinformatics for analysis. For how these disciplines relate, see the overview of the branches of science.

How Epigenetic Regulation Works

Three families of mechanisms account for most of what researchers call epigenetic regulation. They are interdependent rather than separate: a methylated stretch of DNA tends to recruit particular histone-modifying enzymes, and non-coding RNAs can guide both DNA methylation and histone modification to specific places in the genome.

DNA Methylation

DNA methylation is the addition of a methyl group (CH₃) to DNA, most commonly to the 5-position of a cytosine base to form 5-methylcytosine (5mC). In mammals this occurs predominantly at cytosines followed by a guanine, known as CpG sites. The CDC summarises the usual functional effect: typically, methylation turns genes off and demethylation turns genes on. More precisely, methylation of CpG-rich promoter regions is generally associated with stable gene silencing, while methylation within gene bodies has a more complex relationship with expression.

• Writers. DNA methyltransferases (DNMTs) add the methyl group. DNMT1 copies existing methylation patterns onto the new strand after DNA replication (maintenance methylation), whereas DNMT3A and DNMT3B establish new patterns (de novo methylation), especially in early development.
• Erasers. Methylation can be removed. The TET family of enzymes oxidises 5mC, starting a path back to unmethylated cytosine, and methylation can also be lost passively when it is not copied after replication.
• Roles. DNA methylation contributes to genomic imprinting (where only the maternal or paternal copy of a gene is active), to silencing of repetitive and transposable elements, to X-chromosome inactivation in female mammals, and to maintaining cell identity. Abnormal methylation, including silencing of tumour-suppressor genes such as BRCA1 through promoter methylation, is a recurring feature of cancer, and stool-based DNA methylation patterns are used in some colorectal cancer screening tests, as the CDC notes.

Histone Modification

In the nucleus, DNA is wound around protein spools called histones, in repeating units called nucleosomes (each wraps roughly 147 base pairs of DNA around a core of eight histone proteins). The resulting DNA-protein complex is chromatin. How tightly or loosely chromatin is packed affects whether the transcription machinery can reach a gene. The protruding tails of histones can be chemically modified in many ways, including acetylation, methylation, phosphorylation and ubiquitination.

• Acetylation of histone tails is generally associated with open chromatin and active transcription.
• Methylation has effects that depend on which amino acid is modified and how many methyl groups are added. In standard shorthand, H3K4me3 (three methyl groups on lysine 4 of histone H3) marks active promoters, whereas H3K27me3 and H3K9me3 are associated with repression and with condensed heterochromatin.
• Writers, readers and erasers. Enzymes that add marks (for example histone acetyltransferases and methyltransferases), proteins that recognise them, and enzymes that remove them (such as histone deacetylases) together form a regulatory system often described through the “histone code” hypothesis, proposed in 2000, in which combinations of marks are read to produce specific outcomes. The idea is influential, though how strictly marks function as a code remains debated.

Chromatin regulation also includes ATP-dependent remodelling complexes that slide or reposition nucleosomes, and the incorporation of variant histones, which are additional layers beyond covalent modification.

Non-Coding RNA

Most of the human genome is transcribed into RNA, and a large fraction of that RNA does not code for protein. Several classes of these non-coding RNAs take part in epigenetic regulation:

• MicroRNAs (miRNAs) are short RNAs, roughly 22 nucleotides long, that bind to messenger RNAs and reduce their expression after transcription. This post-transcriptional control is related to the RNA interference (RNAi) discovered by Andrew Fire and Craig Mello in 1998, work recognised with the 2006 Nobel Prize in Physiology or Medicine.
• PIWI-interacting RNAs (piRNAs) help silence transposable elements in the germline, protecting genome integrity.
• Long non-coding RNAs (lncRNAs) are longer than about 200 nucleotides and act in diverse ways. The best-known example is XIST, which coats one X chromosome in female mammals and triggers its inactivation, a process that then involves repressive histone marks and DNA methylation.

Whether a given non-coding RNA mechanism counts as “epigenetic” depends on the definition used. Researchers who require heritability through cell division restrict the term to mechanisms that are self-propagating, while others use it more broadly for any regulatory layer above the sequence. Reading a paper critically means checking which definition its authors use.

Epigenome-Wide Studies

An epigenome-wide association study (EWAS) tests many sites across the genome, most commonly DNA methylation, for association with a trait, exposure or disease. It is the epigenetic analogue of a genome-wide association study. Because it tests hundreds of thousands of sites at once, it relies on strict significance thresholds; for the Illumina EPIC array, published work has calculated a 5% family-wise threshold of roughly 9 × 10^-8. For the same multiple-testing logic in genetics, see this site’s guide to genome-wide significance thresholds.

Measurement Platforms

• Methylation arrays are the workhorse of population-scale EWAS. Illumina’s EPIC array interrogates more than 850,000 methylation sites, with greater coverage of enhancers and regulatory regions than its predecessor, the 450K array.
• Bisulfite sequencing treats DNA with sodium bisulfite, which converts unmethylated cytosines to uracil while leaving methylated cytosines intact, so sequencing reveals methylation at single-base resolution. Whole-genome bisulfite sequencing covers the genome; reduced-representation methods cover a CpG-rich subset more cheaply.
• Long-read sequencing platforms can detect some base modifications directly from native DNA without bisulfite conversion. For the platforms themselves, see the comparison of Illumina vs. Nanopore.
• Histone and chromatin profiling uses ChIP-seq (chromatin immunoprecipitation followed by sequencing) with antibodies against specific histone marks or proteins, newer approaches such as CUT&Tag, and ATAC-seq to map open chromatin.
• Expression readouts. To link marks to function, studies pair epigenomic data with transcriptome data; see RNA-seq experimental design and differential gene expression analysis.

Why EWAS Results Are Hard to Interpret

EWAS has well-documented pitfalls, and understanding them is the fastest way to read this literature critically.

• Cell-type heterogeneity. Most EWAS use whole blood, which is a mixture of immune cell types with different methylation profiles. Many diseases and exposures change cell-type proportions, so an apparent methylation signal may simply reflect a different mix of cells. In a published asthma EWAS using the 450K array, over 200 sites were associated with case status, but most lay in granulocyte-specific regions, and nearly all signals disappeared once estimated cell counts were adjusted for.
• Reverse causation. Methylation differences measured in people who already have a disease may be a consequence of the disease, its treatment or its lifestyle correlates, not a cause. Standard EWAS designs cannot separate cause from consequence; approaches such as Mendelian randomization and longitudinal sampling are used to help.
• Technical and demographic confounding. Batch, array position, age, sex, smoking and genetic ancestry all affect methylation and must be modelled.
• Tissue specificity. Blood methylation may not reflect methylation in the brain, liver or other tissue that is biologically relevant to the condition studied, and such tissue is often inaccessible in living people.

Methylation levels at selected sites also change predictably with age, which underlies “epigenetic clocks” that estimate biological age from blood or tissue samples; the first widely used multi-tissue clock was published in 2013. Clocks are used as research tools, and their value as individual clinical predictors is still under study.

Reference Epigenomes

Because each cell type has its own epigenome, the field depends on public reference maps. The NIH Roadmap Epigenomics Program, launched in 2008, produced 111 reference epigenomes of healthy human cells and tissues, mapping DNA methylation, histone modifications, chromatin accessibility and small RNA. The Roadmap program became a founding member of the International Human Epigenome Consortium (IHEC), launched in 2010 with the goal of producing 1,000 reference epigenomes for the international research community.

Common Misconceptions About Epigenetics

• “Epigenetics means you are not stuck with your genes.” Epigenetic marks are real and can respond to environmental inputs, but they are shaped in part by genetic variation, and many are set by developmental programmes rather than lifestyle. The sequence still matters.
• “Epigenetic changes are always inherited by children.” Most epigenetic marks are erased and reset during gametogenesis and early embryonic development. Marks are routinely copied when a body cell divides, which is not the same as passing to the next generation. True transgenerational epigenetic inheritance is well documented in some plants and in model organisms such as worms, but evidence in humans remains limited and contested. Effects of prenatal exposure on a developing fetus, as in the Dutch Hunger Winter, are exposure of the fetus itself, not necessarily inheritance across generations.
• “Epigenetic changes are permanent” or “always reversible.” Both are overstated. Some marks are stable for the life of a cell lineage; others change within months, as the CDC’s smoking example shows. The marks are dynamic, but not freely reprogrammable at will.
• “Epigenetics is a form of Lamarckian inheritance.” It is not an evolutionary mechanism that replaces natural selection. Marks are mostly reset between generations, and the underlying DNA sequence is what natural selection principally acts on.
• “A methylation change in a study is proof that an exposure caused a disease.” As the EWAS section shows, association alone does not establish causation.
• “Thinking positive thoughts or buying a supplement rewrites your epigenome.” Marketing claims of this kind generally go well beyond the evidence. Treat such claims with caution and look for peer-reviewed support in the relevant human outcome.

A Short History

• 1942. Conrad Waddington coined the term “epigenetics” to describe how genes interact with their environment to produce a phenotype during development, a usage tied to epigenesis and cell differentiation.
• 1970s. Robin Holliday and John Pugh (1975), and independently others, proposed that DNA methylation could regulate gene expression and be heritable through cell division.
• 1990s. Molecular work established the DNMT enzymes, genomic imprinting and X-inactivation mechanisms, and RNA interference was described (Fire and Mello, 1998).
• 2000. The histone code hypothesis, associated with Brian Strahl and C. David Allis, framed histone modifications as a readable regulatory language.
• 2000s–2010s. Genome-scale technology made epigenomics possible. The NIH Roadmap Epigenomics Program began in 2008 and IHEC in 2010; methylation arrays enabled large EWAS cohorts.

Who Funds Epigenetics Research

Epigenetics research is funded as part of the broader life-science portfolio rather than through a single dedicated agency.

• NIH. NHGRI funds genomic and epigenomic resource and technology programmes; the National Cancer Institute funds cancer epigenetics; the National Institute of Environmental Health Sciences funds work on environmental exposures and the epigenome; and the National Institute of General Medical Sciences supports basic chromatin and gene-regulation research. Investigator-initiated projects typically use the R01 mechanism. The Roadmap Epigenomics Program was run through the NIH Common Fund, which funds trans-NIH initiatives.
• National Science Foundation. Funds basic chromatin and gene-regulation research in plants and model organisms through its Directorate for Biological Sciences.
• Private and international funders. These include the Howard Hughes Medical Institute, the Wellcome Trust and the European Research Council, and the IHEC consortium coordinates a number of national funders’ epigenome projects.

Journals, Societies and Training

Dedicated journals include Epigenetics, Epigenetics & Chromatin and Clinical Epigenetics, and mainstream genetics titles such as Genome Biology, Nature Genetics and Nucleic Acids Research publish much of the field’s methods and resource work. Epigenetics researchers are commonly active in the broad genetics societies, including the American Society of Human Genetics and the Genetics Society of America, and the international IHEC consortium coordinates reference-epigenome standards.

Training follows the standard life-science route: an undergraduate degree in biology, biochemistry, genetics or a computational field, then a PhD (commonly five to six years in the US) in genetics, molecular biology, cell biology, biochemistry or computational biology, followed by one or more postdoctoral positions. Predoctoral and postdoctoral support commonly comes through NIH fellowships such as the F31 and F32 and institutional T32 training grants. Because EWAS and chromatin-profiling datasets are large, computational skills in statistics and sequence analysis are now part of core training. Epigenetics also supports careers in clinical laboratory diagnostics, biotechnology and pharmaceutical research (where epigenetic drug targets are an active area), and research-administration roles that support genomics programmes.

Epigenetics and Research Administration

Epigenomic studies raise the same administrative questions as other human genomic research, and a few that are specific to the field.

• Data sharing. Human epigenomic datasets are often linked to sequence or phenotype data that can be re-identifying. Funded projects that generate large-scale human genomic data are subject to the NIH Genomic Data Sharing policy, which typically means deposit in controlled-access repositories such as dbGaP. See also the guides on dbGaP submission and controlled-access requests, institutional certification under the GDS policy and submitting sequence data to NCBI SRA, and plan the work in your NIH data management and sharing plan.
• Consent and human-subjects oversight. Informed consent for biobanked samples should anticipate future epigenomic and broad data-sharing uses, which should be reviewed with the institution’s IRB.
• Sample quality and reproducibility. Methylation and expression results depend on input quality, batch design and pre-analytic handling; see the guide on RNA integrity number (RIN) for one common quality metric and the NIH rigor and reproducibility policy for the grant-level expectations. Statistical rigor on EWAS (cell-type adjustment, multiple-testing control, replication) is exactly what reviewers look for.

Frequently Asked Questions

What is epigenetics in simple terms?

It is the study of chemical marks on DNA and on the proteins that package it, which switch genes on or off without changing the DNA sequence. These marks help each cell type use the right genes, and they can respond to age, environment and disease.

What is the difference between genetics and epigenetics?

Genetics concerns the DNA sequence and its inheritance; epigenetics concerns the regulatory layer on top of the sequence that controls how that sequence is used. Both influence traits, and they interact.

What are the main epigenetic mechanisms?

DNA methylation, histone modification and non-coding RNAs, together with chromatin remodelling. They work together and are not independent switches.

Can epigenetic changes be inherited?

Marks are copied when a cell divides, which maintains cell identity within one individual. Most are reset between generations, and inheritance across human generations remains limited and contested, though it is documented in some other organisms.

Can lifestyle change your epigenome?

Yes, some marks respond to exposures. Smoking-related methylation changes are the best-documented example and, according to the CDC, partly reverse after quitting. The size and clinical meaning of most other lifestyle effects are still being established.

What is an EWAS?

An epigenome-wide association study tests epigenetic marks (usually DNA methylation) across the genome for association with a trait or exposure. Its main pitfalls are cell-type confounding and reverse causation.

What degree do you need to work in epigenetics?

Research roles typically require a PhD in genetics, molecular biology, biochemistry or computational biology; laboratory technician and analyst roles are open with a bachelor’s or master’s degree.

Related Guides

This guide is part of a series covering major scientific disciplines from a research-administration perspective. See the overview of the branches of science, and these companion guides: What Is Genetics?, What Is Genomics?, What Is Molecular Biology?, What Is Bioinformatics? and What Is Proteomics?.