The question of whether anxiety is primarily a condition of the mind or of the body has never resolved cleanly, and a major new genetics study published yesterday does not resolve it either. What it does is sharpen the picture of how anxiety’s biological roots reach into territory that clinical psychiatry has not traditionally treated as connected.

We are writers, not clinicians. What follows is a careful reading of the research, not medical or psychological advice.

The study, “Genome-wide analyses of quantitative generalised anxiety symptoms” by Dr. Megan Skelton, Dr. Brittany Mitchell, Professor Thalia Eley, and colleagues from King’s College London, the QIMR Berghofer medical research institute, and the Psychiatric Genomics Consortium, was published in Nature Human Behaviour on 9 June 2026. It is a genome-wide association study, or GWAS: an analysis that searches across the DNA of large numbers of people for genetic variants occurring more often in those reporting higher levels of a given condition. The result is a genetic map, not a diagnosis or a treatment. This is one study, not settled consensus.

What makes this analysis different

Earlier large-scale anxiety genetics studies, including a significant Nature Genetics paper published in February 2026 by Strom and colleagues analysing 122,341 clinically diagnosed anxiety cases, have generally worked with a binary frame: people either have an anxiety disorder or they don’t. The Skelton study takes a different approach, analysing anxiety along a continuous spectrum of symptom severity across 693,869 people of European ancestry drawn from 14 cohorts. The researchers used the Generalised Anxiety Disorder 7-item scale and similar brief self-report measures to quantify how much anxiety each participant reported, then looked for genetic variants that tracked that variation.

The distinction matters because anxiety, as experienced, is not binary. Worry and fear responses exist on a continuum in the population, from ordinary caution to severely disabling disorder. By measuring that spectrum rather than classifying people as cases and controls, the study gains statistical power to identify genetic signals that a binary design might miss. It identified 74 genomic locations associated with anxiety symptom severity, of which 39 had not been identified in any previous analysis. Many of the implicated genes, including PCLO and SORCS3, are particularly active in brain tissue and are involved in how nerve cells communicate with each other.

In terms of sheer participant numbers, this is the largest GWAS of anxiety symptom severity to date. Whether it is the “largest genetic study of anxiety ever conducted” depends on how you count: the Strom et al. study in Nature Genetics has more participants with clinically confirmed diagnoses, and both studies are products of the same wave of large-scale psychiatric genomics. They are complementary rather than competing.

What the correlations with physical conditions mean

The finding most likely to attract misreading is the genetic overlap the study detected between anxiety and several physical conditions. The researchers calculated genetic correlations: a statistical measure of the degree to which the same DNA variants tend to raise the risk of two different traits. For irritable bowel syndrome, the genetic correlation with anxiety was 0.57, a substantial figure. For coronary artery disease, endometriosis, and migraine, the correlations were in the range of 0.20 to 0.27, moderate but statistically significant.

These correlations do not show that anxiety causes bowel problems, or that heart disease makes people anxious, or that the same gene does the same job in the gut and the mind. Dr. Brittany Mitchell, co-first author and Team Head of the Complex Trait Genomics group at QIMR Berghofer, was explicit about this in the King’s College London press release: “Our findings don’t reveal causation or the direction of effect, but they do open up important questions for future research.”

Mitchell adds that shared genetic variants might increase risk for both a physical condition and more severe anxiety symptoms, but it is also true that living with chronic pain or illness can contribute to anxiety. The data cannot distinguish between those possibilities. Both may be operating simultaneously, in different people, for different gene-condition pairs. A genetic correlation of 0.57 between anxiety and IBS is striking, but what it means biologically is a question that still needs answering.

On the gap between the IBS correlation and the others: the magnitude matters. The correlation between anxiety symptoms and IBS (0.57) is roughly comparable to the correlations seen between distinct psychiatric conditions, and substantially higher than the correlations with coronary artery disease and migraine (0.20 to 0.27). The anxiety-gut genetic overlap appears meaningfully larger than the anxiety-heart or anxiety-migraine overlap, a finding that may interest researchers already working on the gut-brain axis but that equally does not, on its own, tell us what to do clinically.

What genetics explains, and what it doesn’t

The study found that common genetic variants explain approximately 6 per cent of the variation in anxiety symptom severity between people. That figure deserves more attention than it usually gets in press coverage of GWAS research, because it reframes what the finding actually says about anxiety’s origins.

Six per cent is not a small number in the context of psychiatric genetics, where capturing this kind of signal from common variants has historically been difficult. But it is also not large in the context of what actually shapes a person’s anxiety. The remaining 94 per cent of variation in anxiety severity is driven by environmental experiences, life events, gene-environment interactions, rare genetic variants not captured in this kind of analysis, and statistical noise. Dr. Megan Skelton, first author on the study, noted directly that “the rise in anxiety rates points to environmental factors, as genetics don’t change much across generations, so reducing anxiety in the population will require these factors to be addressed.”

This is worth holding in mind when reading any coverage that implies the genetic picture is the main story. The genetic mapping is a tool for understanding biological mechanisms; it is not, and the researchers do not claim it is, an explanation for why more people report anxiety now than they did a generation ago.

What the headline overstates

The commissioned headline for this article described findings that researchers say “fundamentally change how anxiety should be understood and treated.” In our reading, no researcher quoted in the primary sources makes that claim. Professor Thalia Eley, lead author and Professor of Developmental Behavioural Genetics at King’s College London, said she hoped the findings would “encourage a new wave of large-scale analyses to accelerate our progress in understanding the genetic architecture of anxiety.” Dr. Skelton described it as “an exciting step forward.” Dr. Mitchell’s comments explicitly disclaimed causation and foregrounded the research questions the study opens, rather than the answers it closes.

The study is a mapping exercise on a scale previously unavailable. It confirms anxiety’s genetic architecture is more complex and more physically entangled than older research captured. It does not propose new treatments. It does not change how clinicians should currently approach anxiety. The finding is worth taking seriously, but it should not be read as the final word.

The more durable contribution may be methodological. By showing that a dimensional, severity-based approach identifies more genetic signal than a binary case-control design, the Skelton study offers a template for future work. The 39 novel loci identified here would not have appeared, or would have appeared less clearly, in a study that only compared diagnosed cases to healthy controls. That is a genuine advance in how this kind of research gets done.

Limitations worth noting

The study is conducted exclusively in people of European ancestry. The researchers are candid about this: polygenic risk scores built from European data explained only 1.2 to 2.9 per cent of anxiety variance when applied to African and South Asian ancestry samples, and the authors note that ancestry-specific analyses are necessary to understand population-specific genetic risk. At present there are not enough datasets with both anxiety measures and genetic data from African or South Asian populations to run statistically meaningful GWAS. The gap is a known problem in psychiatric genetics, and the Skelton study does not close it.

The study also relies on self-reported anxiety symptoms. Self-report scales like the GAD-7 measure how people describe their experience, not a biological state directly. Correlations identified via self-reported symptoms may not map cleanly onto biological processes, though the researchers argue, with support from prior work, that symptom-severity measures are genetically informative.

If you are experiencing anxiety that is persistent, severe, or affecting your daily life, a GP or mental health professional will be more useful than any study of population-level genetics. This finding does not tell you anything specific about your own risk or your own situation.

What to watch next

The Skelton study arrives alongside, not instead of, the earlier Strom et al. work in Nature Genetics. Between them, 2026 has produced two of the largest and most methodologically varied anxiety GWAS analyses ever published, identifying dozens of loci that had not previously been mapped. Whether those loci translate into any concrete diagnostic or therapeutic advance will depend on subsequent research into what the implicated genes actually do in the brain and body, and whether any of the genetic overlaps with physical conditions can be shown to involve a shared biological mechanism rather than shared environmental exposure.

The researchers indicate they intend to publish further analyses, and the Psychiatric Genomics Consortium, which coordinated this work, continues to expand its datasets. The next wave of studies will need to move beyond European ancestry samples if the findings are to be generalisable. On the question of gut-anxiety overlap specifically, the correlation with IBS at 0.57 is high enough that it may attract targeted follow-up work. Whether that follows the genetics, the microbiome, or both will be worth watching.