PITTSBURGH
Ever wondered why the antidepressant that worked wonders for your friend did nothing for you? Or why one person sails through a statin while another is laid low by aching muscles? A landmark American study suggests a big part of the answer may be written in your DNA, and that almost nobody is immune.
Researchers using data from the National Institutes of Health's All of Us Research Program have found that every single participant in their analysis of nearly 100,000 people carried at least one genetic variant in genes known to shape how the body processes medicines. Along the way they uncovered thousands of variants that had never been described before.
The study, published Wednesday, November 11, 2026, in the journal Science Translational Medicine, also found that roughly one in four participants had actually taken a medication that these variants could affect.
The drugs involved aren't obscure. They include SSRI antidepressants, cholesterol-lowering statins, cancer chemotherapy, and the widely used blood thinner Plavix (clopidogrel), according to STAT, which first reported the findings.
Genes That Decide How Your Pills Behave
The field behind the study is called pharmacogenomics, which is a mouthful for a simple idea: your genes can change how a drug works in your body.
Many medicines are broken down, activated, or moved around the body by enzymes and transporter proteins. The instructions for building those proteins come from specific genes, known as pharmacogenes. If you carry a variant that makes an enzyme work faster, a drug might be cleared before it can do its job. If the enzyme works more slowly, the drug can build up to levels that cause side effects. Some drugs, including Plavix, need to be activated by an enzyme before they work at all, so a weak version of that enzyme can leave a patient with less protection than their doctor expects.
The new study zeroed in on 16 well-established pharmacogenes. The researchers' earlier preprint version of the work listed genes such as CYP2C19, which affects Plavix and some antidepressants; SLCO1B1, linked to statin side effects; DPYD, which is tied to severe reactions to certain chemotherapy drugs; and VKORC1 and CYP2C9, which affect the blood thinner warfarin. The headline finding was blunt. Everyone in the study carried at least one variant in these genes.
Why Diversity Made the Difference
What makes this study stand out is who was in it. All of Us was built on purpose to enroll Americans who have historically been left out of medical research, and that design paid off.
According to STAT, the program's diverse participants allowed the team to identify variants that had not turned up in earlier work using resources such as the U.K. Biobank, whose participants are mostly of European ancestry. That matters, because much of what doctors know about drug-gene interactions was learned from narrower populations.
A 2024 preprint version of the analysis, posted on medRxiv, showed the scale of those gaps. Among 98,590 participants with whole-genome sequencing, about half were of European ancestry, 23% of African ancestry, and 16% of admixed American ancestry. When the researchers compared their results with standard reference data used to guide prescribing, roughly a quarter of the allele frequencies and more than a third of predicted phenotype frequencies did not match. The biggest mismatches were in African and admixed American groups.
The preprint also reported striking differences between groups. For example, an actionable variation in the gene CYP3A5, which affects certain transplant drugs, was found in about 72% of participants of African ancestry versus about 13% of those of European ancestry. The team reported 491 allele frequencies and 129 phenotype frequencies for specific ancestry groups for the first time. (The final, peer-reviewed paper may report updated figures.) In plain terms: if your family roots are outside Europe, the prescribing guidance your doctor relies on may have been built with less information about people like you.
One in Four Has Already Been Exposed
The study didn't just look at DNA. It linked genetic data with participants' medical records, which let researchers see who had actually been prescribed the affected drugs.
About one in four participants had taken a medication that their own variants could affect, according to STAT. The earlier preprint found that the most common examples included the heartburn drugs omeprazole and pantoprazole and the statin atorvastatin, medicines found in millions of American households.
The researchers caution that the true number may be even higher, because medical records often don't capture every prescription a person has filled and because some important genes weren't included in the analysis.
Consider Plavix as an example of why this matters. The drug is prescribed to many heart attack and stent patients to keep dangerous clots from forming. But it is a "prodrug," which means the body has to convert it into its active form, and that job depends heavily on the CYP2C19 enzyme. People who carry low-function versions of the CYP2C19 gene may convert less of the drug, which can leave them with less protection than their cardiologist intended. The FDA already carries a boxed warning on Plavix about reduced effectiveness in such "poor metabolizers," and the new study makes clear that variation in genes like this one is the norm rather than the exception.
Statins tell a similar story from the other direction. Certain variants in SLCO1B1 reduce the liver's ability to take up some statins, raising blood levels of the drug and the risk of muscle pain and damage. For a patient who quits a statin because of side effects, a genetic explanation could point toward a different drug or dose rather than abandoning cholesterol treatment altogether.
Expert Perspective
Senior author Philip Empey, who directs the Center for Clinical Pharmaceutical Sciences at the University of Pittsburgh, said the results point to a need for more attention to genetics in everyday clinical care.
But he was careful not to oversell it. If someone has had a bad experience with a medication, "we can't say it's absolutely because of genetics," he told STAT. Genetics can, however, point to where differences exist. Age, kidney and liver function, other medicines, diet and many other factors also shape how a drug works.
Loren Saulsberry, a health policy researcher who was not involved in the study, said the work suggests that "genetically guided clinical care adds benefit across a diverse population."
There are still real hurdles. Most people have never had their genomes sequenced, and many physicians don't know how a given variant affects the drugs they prescribe. Many variants can affect a single drug, which makes interpreting results complicated. Turning genetic findings into routine care remains a work in progress.
The Funding Cliff Hanging Over the Science
The study arrives at an awkward moment for the program that made it possible.
All of Us, launched with a goal of enrolling 1 million Americans, is facing a steep drop in funding. STAT reported that its money from the 21st Century Cures Act ends in fiscal year 2027, and the program expects about $122 million in appropriations once a budget is finalized, roughly 80% below its 2023 peak.
The squeeze is already being felt. The program has stopped returning genetic test results to participants, and about 5,000 enrolled children have data that aren't yet available to researchers. Joshua Denny, who led All of Us as CEO until the end of October and is now at the University of Chicago, told STAT that engagement activities, including enrollment at academic medical centers, federally qualified health centers and mobile units, "were cut back dramatically." Sara Van Driest, director of pediatrics for All of Us, said the program will need "stable funding in the years ahead."
Supporters warn that slowing enrollment risks undermining exactly what made this study possible: a large, diverse population that reflects the whole country, not just part of it.
What This Means for You
You don't need a genome sequence to benefit from this research, but it's worth knowing that your genes may be part of the story if a medicine isn't working as expected. Some practical steps:
The Bottom Line
The idea that one dose fits all has always been a convenient simplification. This study makes it harder to defend. Every person in the analysis carried genetic differences that could change how their body handles common drugs, and about a quarter had already taken a medication those differences might affect.
Researchers stress that genetics won't explain every bad reaction. But the work moves medicine one step closer to a future where prescriptions are tailored to the person taking them, including people whose ancestry has long been underrepresented in research. Whether that future arrives quickly may depend as much on Congress and research budgets as on the science itself.
