Summary: For decades, people have blamed inherited “bad genes” for much of their health risk. A major new study now shows that the environments we live in—the full mix of what we eat, breathe, and encounter over a lifetime—can be equally influential. Researchers re-analyzed two decades of U.S. health data, testing more than 115,000 potential links between 619 environmental exposures and 305 health measures to map how the exposome shapes disease risk.
The team found that individual exposures, such as a single pollutant or nutrient, usually account for only a small portion of differences in health. But when multiple exposures are considered together, their combined effect—the exposome—can explain as much variation in some traits as important genetic variants do.
Key Facts
- Aggregate effect: Most single environmental measures explain under 1% of the differences in clinical outcomes. Examining up to 20 exposures at once raises that explanatory power to levels similar to some genetic predictors.
- Triglycerides example: A specific combination of 20 factors—among them trans fats, polychlorinated biphenyls (PCBs), and vitamin E forms—accounted for about 43% of the variation in triglyceride levels, a key heart-disease risk marker.
- Adapting GWAS methods: The researchers applied a genome-wide association study (GWAS)-style approach to the exposome, shifting away from the conventional one-exposure-at-a-time studies.
- No single “smoking gun”: As lead author Chirag Patel explains, there is seldom one dominant toxic agent; instead, many exposures each contribute modestly, and combinations can be especially consequential.
- Open resource: The team released an open Phenome-Exposure Atlas so other scientists can explore and build on these findings.
Source: Harvard
Background
Genetic research has long helped explain why people vary in their susceptibility to conditions like cancer, diabetes, and heart disease. But genes tell only part of the story. The other part is the exposome—the full set of non-genetic influences, including diet, pollutants, infections, medications, and lifestyle factors—that accumulates over a lifetime and interacts with biology to affect health.
A team led by researchers at Harvard Medical School has produced one of the largest systematic analyses to date of exposome–health associations. Using 20 years of data from the U.S. National Health and Nutrition Examination Survey (NHANES), they tested more than 115,000 exposure–outcome pairs and identified thousands of statistically robust links, showing the value of studying exposures in aggregate rather than in isolation.
While many individual exposures had only modest associations with health outcomes, models that combined multiple exposures generally explained substantially more variation. “A single exposure might not move the needle much, but the combined soup of exposures can be as powerful as your DNA in shaping disease risk,” said Chirag Patel, associate professor of biomedical informatics.
Published in Nature Medicine, the study highlights how existing large-scale public datasets can be repurposed to map exposomic influences and identifies priorities for future, more targeted research.
Study approach
Patel and co-leader Arjun (Raj) Manrai, both in biomedical informatics at HMS, built their approach by adapting techniques from genetics. Rather than testing one exposure and one outcome at a time, they implemented an exposome-wide association strategy analogous to GWAS, scanning many exposures against many phenotypes to create a comprehensive catalog of associations.
They partnered with meta-research expert John Ioannidis (Stanford) and colleagues from collaborative initiatives including the Network for Exposomics in the United States and the Human Exposome Project to scale the analysis and ensure rigor.
Key results
Analyzing ten waves of NHANES, the researchers compared 619 exposure indicators to 305 quantitative health measures—ranging from body mass index and blood glucose to lung function and lipid profiles. More than 5,600 associations were statistically significant and reproducible across waves.
Across many phenotypes, individual exposures explained less than 1% of between-person variation. But when up to 20 exposures were modeled jointly, the average explanatory power rose to about 3.5% for 120 outcomes—comparable to some single genetic variants. Some exposure combinations showed especially large effects: the 20-factor mix tied to triglyceride variation accounted for roughly 43% of differences in that marker.
Still, the authors stress that most combinations explain only modest amounts of variability and that exposures tend to act in interconnected ways rather than in isolation, which complicates causal interpretation.
Next steps and applications
The study serves as a reference point for future exposome research. The team plans to expand analyses to include additional exposures and outcomes, study how early-life environmental signals influence later disease, and explore how exposomic measures could be incorporated into clinical risk assessment tools.
To encourage follow-up work, the researchers have made their data and code publicly available in the Phenome-Exposure Atlas of Health and Disease Risk as a resource for hypothesis generation and prioritization of mechanistic studies. “Large-scale analyses help us figure out where to zoom in,” Manrai said—pointing to the need for detailed follow-up to establish causation.
The authors also envision future integration of exposomic insights with wearable devices and artificial intelligence so people and clinicians can receive real-time, personalized guidance about how current exposures may influence short-term and long-term health.
Funding and disclosures
This research was supported by the National Institutes of Health (grants R01ES032470, R01DK137993, U24ES036819), the U.S. Department of Agriculture, and the Office of Naval Research (N000142412687).
Key Questions Answered:
A: Partly, yes. Unlike DNA, many components of the exposome are modifiable. Although you can’t control every environmental factor, identifying harmful combinations—such as certain fats paired with specific pollutants—may allow targeted lifestyle changes that help offset genetic risk.
A: Not quite. The study undermines the myth of a single dangerous ingredient. It shows single factors rarely have a dramatic impact alone; it is the cumulative mixture and interactions across diet, air quality, stress, and other exposures that often matter more.
A: The authors imagine that future systems could combine wearable health data with environmental sensors and AI to deliver personalized, real-time advice—for example, suggesting dietary adjustments on high-pollution days to reduce acute cardiovascular stress.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- Journal paper was reviewed in full.
- Additional context added by staff.
About this genetics research news
Author: Katie Brace
Source: Harvard
Contact: Katie Brace – Harvard
Image: The image is credited to Neuroscience News
Original Research: Open access. “An atlas of exposome–phenome associations in health and disease risk” by Chirag J. Patel, John P. A. Ioannidis & Arjun K. Manrai. Nature Medicine
DOI: 10.1038/s41591-026-04266-0
Abstract
An atlas of exposome–phenome associations in health and disease risk
Non-genetic exposures that make up the exposome—diet, lifestyle, infections, pollutants and other environmental factors—shape many clinical traits, yet evidence has been fragmented. This study performed an exposome-wide association analysis using 619 exposure indicators and 305 quantitative phenotypes across ten NHANES waves. Replicable signals clustered most in cardiometabolic and anthropometric traits, linking nutrient biomarkers and lipophilic pollutants with BMI, glycated hemoglobin, and lipid measures. Triglycerides showed the strongest ties to multidomain exposures, notably trans fatty acids, persistent pollutants, and vitamin E isoforms. For pulmonary traits, tobacco-specific and carcinogen biomarkers aligned more closely with reduced lung function than short-lived nicotine metabolites. While single exposures generally had modest effects, combined poly-exposomic models explained phenotypic variation on par with genome-wide polygenic scores. Exposome “globes” reveal an interconnected architecture where exposures rarely act alone, complicating causal claims but offering a broader view of environmental risk. These findings point to which exposures may enhance disease risk assessment, population surveillance, and prioritization for longitudinal exposomics.