For years, scientists have focused on DNA as the main driver of health – the tiny genetic differences that shape our risk of disease. But everyday experience tells a different story. Two people with similar genes can end up on completely different health paths.
That difference comes down to everything else: the air you breathe, the food you eat, the stress you carry, and the environment you live in.
Scientists call this full lifetime mix of influences the “exposome.” Now, a new study takes one of the most detailed looks yet at how much it really matters.
Health risks come in layers
In the study conducted by Harvard Medical School (HMS), researchers tested more than 115,000 exposure–health links using decades of U.S. survey data.
The results revealed a clear pattern. On their own, single exposures don’t explain much. But when they stack together, their combined impact starts to look much more like genetics.
“While one single exposure might not make a massive difference in your health, the cumulative soup of exposures can be just as powerful as your DNA in determining your risk of certain diseases,” said first author Chirag Patel, an associate professor of biomedical informatics at HMS.
Health headlines don’t always agree
Environmental health research often falls into a familiar trap: scientists study one ingredient, one pollutant, or one nutrient, then argue endlessly about whether it’s harmful, harmless, or both.
Patel and senior author Arjun (Raj) Manrai say this piecemeal approach helps explain why the public gets whiplash from health headlines.
You might see one study claiming something in your pantry is dangerous and another saying it’s fine or even beneficial.
The truth may be that the effect is small, context dependent, and easier to detect when you look at combinations rather than isolated factors.
“To date, the field has been imbalanced. More has been done on the side of using genetics to individualize treatment, and not much has been done on the side of environmental exposures,” Patel said.
“We really wanted to build a robust, large-scale compendium of these associations for the exposome,” Manrai added.
To do that, they leaned on an approach genetics has used for years: scan everything in a systematic way, then follow up on the strongest signals. That’s what genome-wide association studies (GWAS) do for DNA. Patel and Manrai wanted the exposome equivalent.
Where this data comes from
Instead of launching a brand-new, expensive mega-study, the researchers used a treasure trove that already exists: the National Health and Nutrition Examination Survey (NHANES), which is run annually by the U.S. Centers for Disease Control and Prevention.
They pulled 20 years of NHANES data and tested 619 exposures – alongside DNA-related influences – across 305 health outcomes (such as BMI, blood sugar, lung function, and more).
That created more than 115,000 possible associations, and they found over 5,600 that were statistically significant.
The study doesn’t claim these associations prove causation – NHANES is observational – but it provides a way to map the terrain and identify where the strongest signals lie.
It’s the combination that matters
Here’s what they found when they asked, “How much does this explain?” – meaning how much of the difference between people’s health outcomes could be statistically linked to exposures.
Across hundreds of outcomes, single exposures tended to explain less than one percent of the variation. This may sound disappointing, but most complex diseases arise from many small influences, not one giant lever.
Then they did something closer to real life: they looked at multiple exposures at the same time, with up to 20 exposures per outcome.
That raised the explanatory power to an average of 3.5 percent across 120 outcomes, which the team notes is in the same ballpark as the effect sizes of some individual genetic variants.
In other words, exposures often look weak when you isolate them, but they become more meaningful when you treat the exposome as it actually exists – as a cluster.
“On the whole, there’s no smoking gun; every exposure seems to matter a little bit, and exposures matter more when you consider them in aggregate,” Patel said.
Some exposures matter more
Most exposure bundles still explained only modest differences, but a few stood out. One of the strongest examples involved a particular mix of 20 exposures.
These included trans fats, pollutants like polychlorinated biphenyls (PCBs), and vitamin E levels. Together, they explained 43 percent of the variation in people’s triglyceride levels. Triglycerides are a well-known risk factor for heart disease, so that’s not a trivial link.
The researchers stress that this kind of high explanatory power wasn’t the norm. But it’s proof that, at least in some cases, exposures can add up to something big enough to matter clinically.
Building a health research roadmap
The team sees this work as a starting map, not a final answer key. Large-scale scans are great for spotting patterns, but they can’t tell you the biological mechanism.
They also can’t determine whether an exposure causes a health outcome or simply correlates with another factor, including DNA.
That’s why the researchers are framing the study as a jumping-off point – a place to generate hypotheses, then design focused studies to test causality.
“Large-scale analyses like this are an agnostic, systematic way of generating hypotheses, but we then need to conduct detailed mechanistic evaluations of exposures and their associations with disease to determine causation,” Manrai said. “We’re zooming out to figure out where to zoom back in.”
To make that possible, they’ve published the results and tools in an online resource called the Phenome-Exposure Atlas of Health and Disease Risk, so other researchers can dig into specific associations.
A smarter way to track health
Patel and Manrai plan to expand this approach to include more exposures and outcomes, with a focus on how early life exposures shape disease risk decades later.
They also hint at something more futuristic but not totally far-fetched. Many people already wear devices that track sleep, heart rate, activity, and sometimes even glucose.
The authors imagine layering exposure information into this kind of monitoring, giving people a clearer real-time sense of what might be affecting their health.
“It may seem like a pie-in-the-sky, Star Trek-y vision, but I see a future where exposomic information is integrated into these systems with AI so that a person can understand, in real time, how exposures may be impacting their day-to-day or even hour-to-hour health,” Patel said.
Genetics still matter. But this study is a reminder that DNA isn’t destiny when it comes to health and that the environment isn’t one thing, either. It’s a constantly shifting pile of inputs that can stack, interact, and compound over time.
If medicine wants to get serious about preventing disease (not just treating it once it’s obvious), mapping that pile may be one of the most practical next steps.
The study is published in the journal Nature Medicine.
—–
Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates.
Check us out on EarthSnap, a free app brought to you by Eric Ralls and Earth.com.
—–
