The study, led by Sophie van Baalen, Thomas Verra and Michelle Habets of the Rathenau Instituut in the Netherlands, together with colleagues at Wageningen University and Erasmus MC, conducted nineteen semi-structured interviews with scientists working in academia and industry between September 2023 and January 2024. Fifteen of the respondents were academic researchers and four worked for commercial companies, with expertise spanning biomedical research, plant science and microbiology. Using snowball sampling and thematic analysis with Atlas.ti software, the team captured the views of researchers across ten European countries and one respondent from the United States, continuing recruitment until no new themes emerged from the transcripts.

The technical logic behind epigenetic editing explains much of its appeal. Unlike CRISPR-Cas9 gene editing, which slices both strands of the DNA double helix to alter the genetic sequence, epigenetic editing tools such as CRISPR-dCas9 and zinc finger proteins bind to specific DNA sequences without cutting them. Instead of changing the letters of the genetic code, they attach or remove chemical marks on the DNA and its associated proteins, dialing gene activity up or down. Because no DNA breaks are introduced, researchers assume the risk of genomic instability is dramatically reduced. Off-target effects, a persistent worry in gene editing, are viewed as less dangerous in the epigenetic version because misplaced edits do not coincide with strand cuts and may fade over time as the cell’s own machinery reverses the marks.

Respondents also described epigenetic editing as fundamentally more subtle than gene editing. Where gene editing acts like a binary switch, introducing or eliminating genetic functions outright, epigenetic editing was compared to a thermostat that fine-tunes the volume of gene expression. Because cells naturally modify their epigenome constantly as part of ordinary biology, scientists framed the technique as working with, rather than against, the cell’s native processes. Several researchers contrasted this precision favorably with epigenetic drugs, a class of therapeutics that targets epigenetic enzymes broadly and is notorious for lacking specificity. One interviewee emphasized the absence of the translocation problems and genetic instability that plague strand-cutting technologies, noting that unlike base editing, prime editing or conventional gene editing, epigenetic editing never cuts the DNA at all.

Yet the same scientists who endorsed this dominant vision simultaneously questioned its foundations, sometimes without being prompted. The reservations clustered around three scientific uncertainties: reversibility, heritability and complexity. On reversibility, researchers acknowledged that while a limited number of studies have shown epigenetic edits can be reversed or remain stable, it is currently impossible to predict whether an edit at a particular genomic location will persist or vanish, and for how long. This creates an awkward paradox, because the stability needed for durable medical treatments and agricultural applications is exactly what undermines the promised safety net of reversibility. As one respondent explained, scientists do not really know what makes some epigenetic modifications stick around for years while others disappear within days, and following patients long enough to find out would require decades-long cohort studies.

Heritability raised equally thorny questions. For epigenetic editing to work in medicine or agriculture, edits must survive cell division, a property called mitotic heritability, and respondents disagreed about whether they reliably do. The possibility of intergenerational and transgenerational inheritance troubled some biomedical researchers, who worried about unintended effects on the offspring of treated patients, while some plant scientists actually counted on epigenetic edits fading out over generations, reasoning that edited crops escaping into the wild would lose their modifications naturally. The third concern, complexity, struck at the heart of the technology’s predictability. Gene expression is governed by intertwined networks in which cause and effect are not linear; altering an epigenetic mark at one location can trigger cascades of unforeseen changes across hundreds of other genes, especially when the three-dimensional folding of DNA and crosstalk between cells enter the picture. One respondent noted that outside of genomic imprinting, they could not think of a single epigenetic network that is well enough characterized to guarantee a clear therapeutic output.

Despite these doubts, three distinct visions of the technology’s medical future emerged among respondents. The prevailing outlook was hopeful but modest: epigenetic editing could eventually treat cancers driven by epigenetic changes, boost the effectiveness of CAR-T immunotherapies, and address rare diseases caused by epimutations, with early clinical trials restricted to patients who have exhausted all other options. Some argued that medicine routinely advances without fully understanding a drug’s mechanism, so demanding perfection before testing would mean never developing the therapy at all. At the cautious extreme, a minority, particularly basic researchers outside translational work, warned that epigenetic editing could prove less safe than gene editing, since introducing epimutations might reactivate dormant transposons or derail cellular identity in ways scientists can no longer control once edited cells are inside the body. At the opposite pole, a small group envisioned a medical revolution in which epigenetic editors retune multiple genes simultaneously, potentially transforming treatment of autoimmune disease, diabetes, Alzheimer’s and even aging, with speculative applications ranging from skin-rejuvenating creams to cures for HIV and chronic hepatitis B.

The agricultural picture diverged sharply. Plant scientists interviewed for the study did not consider epigenetic editing a commercially viable breeding technology, largely because seed companies require traits that remain stable across generations and environmental conditions, from the controlled greenhouse to the unpredictable open field. Backcrossing can eliminate unwanted off-target changes in plants, removing one of epigenetic editing’s main selling points. Still, respondents sketched hypothetical applications that could eventually prove transformative, such as plant varieties that switch on disease-resistance genes only when a pathogen is actually present, or epigenetic control of flowering, which could dramatically shorten breeding cycles for seed production.

The study’s timing makes its findings particularly pointed. While most respondents questioned whether the technology is ready for real-world deployment, companies are already racing ahead: OMEGA Therapeutics completed a first-in-human clinical trial using epigenetic editing to suppress the oncogene c-MYC in twenty-four participants before filing for bankruptcy in early 2025, and Tune Therapeutics is currently recruiting patients for a trial of an epigenetic silencing therapy for chronic hepatitis B. The authors highlight a mismatch between the private sector’s focus on common, profitable conditions such as high cholesterol and obesity, exemplified by celebrated preclinical results showing durable cholesterol reduction in mice and primates, and the academic community’s caution. They argue that the dominant safety narrative is performing rhetorical work, positioning epigenetic editing as publicly acceptable in ways that may prove premature, and warn of a hype-disappointment cycle reminiscent of the gene therapy backlash of the 1990s. The researchers call for explicit reflexivity within the field, public engagement and citizen participation in shaping the technology’s future, and urge scientists to clearly articulate what evidence is truly needed before epigenetic applications move toward the clinic, arguing that making these visions explicit allows scientists, policymakers and the public to reflect on, adapt and co-create the trajectory of this emerging technology rather than simply inherit whatever future the loudest promises deliver.

Subject of Research: Scientific visions, expectations and reservations regarding epigenetic editing and its responsible innovation

Article Title: Visions, expectations, and reservations in epigenetic editing: towards responsible innovation

Article References: van Baalen, S., Verra, T., Macnaghten, P., Bunnik, E., & Habets, M. G. (2026). Visions, expectations, and reservations in epigenetic editing: towards responsible innovation. Epigenetics Communications, 6(1), Article 6. https://doi.org/10.1186/s43682-026-00047-5

Image Credits: AI Generated

DOI: 10.1186/s43682-026-00047-5

Keywords: epigenetic editing, epigenome, CRISPR-dCas9, gene expression, responsible innovation, gene editing, biotechnology, clinical trials, plant breeding, science and technology studies, research ethics, public acceptance