The epigenetics studies the chemical modifications that regulate gene activity without altering the DNA sequence. An emerging field, epigenomics, reveals how our lifestyles, diets, and environmental exposures modulate genetic expression and affect health. One of the most fascinating areas is the so-called epigenetic noise, a phenomenon that allows cells to temporarily change their identity.

Epigenetic noise and immune tolerance

A recent study by Andrew Koh's group at the University of Chicago, published in Nature in August 2025, shows that the medullary epithelial cells of the thymus (mTECs) adopt a looser, “noisier” chromatin state to express genes from other tissues and thus train the immune system【599630230769128†L150-L214】. This epigenetic flexibility allows mTECs to display proteins from different organs to developing T cells, thereby eliminating autoreactive cells and promoting tolerance.

Interestingly, the transition to this noisy state is associated with the suppression of the protein p53, known as the “guardian of the genome.” When the researchers artificially activated p53, the chromatin stabilized, the noise disappeared, and the mTECs stopped displaying genes from other tissues, causing autoimmune disease【599630230769128†L218-L233】. This suggests that epigenetic “noise” is a physiological mechanism that balances cellular stability with the need to adapt and educate the immune system.

The study also found that deleting p53 in lung cancer cells increases epigenetic noise, allowing the tumor cells to reprogram their identity and become more aggressive. [599630230769128†L244-L247] Thus, understanding how epigenetic noise is modulated could open up avenues for new regenerative and oncology therapies.

Transposons and epigenetics in the blood

Another interesting line of research focuses on the transposons, DNA fragments that can “jump” within the genome. SVA elements, for example, act as enhancers and present bivalent epigenetic marks (H3K9me3 and H3K27ac) that regulate genes involved in the maturation of red and white blood cells【227440307021256†L149-L163】. These discoveries show that epigenetics is not a passive process, but a central player in cellular differentiation.

PromoterAI: Artificial intelligence for decoding non-coding variants

In addition to epigenetics, artificial intelligence is transforming genomics. Promoters are regions of DNA that control when and where a gene is activated. Mutations in these promoters can alter expression and cause rare diseases, but their interpretation is complex because they occur outside of coding regions. To address this challenge, Illumina researchers developed PromoterAI, a deep learning algorithm that predicts how variants in promoters affect gene expression【610559585256559†L182-L201】.

PromoterAI was trained with chromatin accessibility, histone modifications, and transcription factor binding data from tens of thousands of genomes and validated with information from the Genotype-Tissue Expression (GTEx) and of the UK BiobankIn tests, the model identified promoter variants that dramatically altered protein expression and explained up to 6% of genetic diseases; when combined with splice variants, the figure increases to 20% of genetic diseases.

This breakthrough demonstrates how algorithms can uncover mutations that traditional medicine overlooks. When integrated into clinical practice, tools like PromoterAI could help diagnose rare diseases, predict treatment response, and guide gene-editing therapies. They also complement personalized nutrition by identifying variants that affect nutrient metabolism and inflammatory response.

Implications for precision nutrition

The findings on epigenetic noise and non-coding variants have implications for personalized nutrition:

• Metabolic flexibilityThe ability of cells to change their identity suggests that dietary interventions and exercise can reprogram tissues such as the liver or muscles. Restricting calories, practicing intermittent fasting, or modifying macronutrient composition could influence the epigenetics of metabolic cells.
• Supplements and nutraceuticalsSome bioactive compounds (e.g., polyphenols, curcumin, resveratrol) modulate the activity of p53 and epigenetic enzymes. Understanding these interactions helps design supplements that promote epigenetic homeostasis.
• Diagnosis of intolerancesIdentifying variants in promoters associated with digestive enzymes or nutrient transporters could explain why some people respond poorly to certain foods. AI algorithms can detect these variants and suggest dietary changes.
• Holistic approach: Interventions should not focus solely on genetics. The microbiome, the environment, and emotional state must also be considered. Mefood Omics and Alimentomics We investigate how to integrate epigenetic and microbiota data to adjust the nutritional plan.

Prudence and education

While the prospects are exciting, we are still far from widespread application of these findings. Epigenetics is dynamic and sensitive to factors such as diet, stress, toxins, and age. AI algorithms require validation in diverse cohorts and access to quality clinical data. Furthermore, there is a risk of generating undue expectations or misusing the information.

To move forward ethically and responsibly we must:

• Promote research and diversity: Epigenetics and non-coding variants need to be studied in diverse populations to improve equity.
• Protect privacy: As with PGS, epigenomics research must ensure consent and legitimate use of data.
• Training professionals and citizensNutritionists, physicians, and bioinformaticians need to update their knowledge of epigenetics and AI. Users, for their part, need to understand the limits of science.
• Integrate the evidence: Combining genetics, epigenetics, microbiota, habits and clinical data to design practical recommendations.

On our platforms Oorenji and Mefood Omics We work with models that respect these premises. Our goal is to translate scientific findings into evidence-based nutrition and wellness plans, without promising miracle solutions.

Conclusion: a promising but complex future

Epigenetics and artificial intelligence are opening unexpected doors in personalized medicine. Epigenetic noise reminds us that cells are not static entities: they can be flexible when the body needs it, but that flexibility can also be exploited by cancer. Algorithms like PromoterAI allow us to see parts of the genome that until now were invisible to clinical use. However, we must proceed with caution: science is advancing rapidly, but practical application must follow ethical criteria and be based on solid studies.

For those who want to improve their health, the key remains adopting balanced habits and seeking professional support. Exploring your epigenome and non-coding variants can offer valuable insights, but it's no substitute for a healthy diet, regular exercise, and plenty of rest. Mefood Omics, Oorenji and Alimentomics We help you integrate the best of science into your daily life.

References

1. Epigenetic noise: Unappreciated process helps cells change identity – University of Chicago News
2. Thymic epithelial cells amplify epigenetic noise to promote immune tolerance – Nature
3. Bivalent chromatin state of composite transposons in hematopoiesis – Nature Genetics
4. AI tool identifies disease-driving promoter mutations – The Scientist
5. Predicting expression-altering promoter mutations with deep learning – Science