Your cells keep a record of every drink. Not in your DNA sequence - that doesn't change. The record is written in methyl groups: tiny chemical tags that stick to DNA and change how genes get read, without touching the code underneath. A new study of 13,970 people maps where alcohol leaves those tags, in more detail than anyone has managed before.

What they did

This may be the largest methylome-wide association study (MWAS) of alcohol so far. An MWAS works like a genome-wide scan. Instead of hunting genetic variants, it sweeps every methyl tag in the genome - the methylome - looking for the ones that move with a given behavior. 13,970 people. Blood as the starting material.

Whole blood is a mix of many cell types, so the team ran epigenomic deconvolution to pull the signals apart across 12 separate blood cell populations: neutrophils, several kinds of T-cells, eosinophils, and others. Then they checked the whole-blood results against a separately published alcohol MWAS to see what replicated.

What they found

  • Whole blood: 1,266 sites across the methylome hit statistical significance. That is a wide footprint. Alcohol's epigenomic effects don't sit in a handful of genes.
  • The top replicating signal landed in SLC7A11, a gene that had already shown up in earlier alcohol methylation work.
  • Cell-type-specific signals were much thinner: eight associations each in neutrophils and CD8+ naive T-cells, three each in CD8+ memory T-cells and eosinophils, one in T regulatory cells.
  • Strongest cell-type hit: PDIA5, in CD8+ naive T-cells.
  • Whole-blood findings overlapped with genome-wide association studies (GWAS) of problematic alcohol use. They did not overlap with GWAS of alcohol consumption in general.

What it means

That last split is the interesting part. These methylation marks line up with harmful drinking patterns, the kind GWAS of alcohol use disorder picks up. They don't line up with how much a person drinks overall. Volume isn't what the signals are reading.

The team also flagged a molecular pathway that hadn't come up in earlier alcohol methylation work: Rho GTPase signaling, which handles immune regulation and cell structure. Worth digging into, and possibly a target for treating problematic drinking.

Direction is the open question. The changes could be residue of long alcohol exposure, marks laid down over years. Or they could partly reflect a biological tendency that was there first. The researchers lean toward exposure, partly because the signals lined up consistently across several cell types. Their study design can't settle it.

Caveats

Cross-sectional design. Everything was measured at a single point in time, so nothing here shows whether the methylation changes came before the drinking or after it. The signals are blood-specific, and blood isn't brain, liver, or the other tissues alcohol hits more directly. In some cell-type analyses a single association cleared significance, which says statistical power at that resolution is thin.

Source: Molecular Psychiatry, 10.1038/s41380-026-03477-8