A polygenic risk score (PRS) estimates how your combined common DNA variants shift your chance of a complex condition, such as coronary artery disease or type 2 diabetes, compared with other people. It is a relative ranking, not a diagnosis and not a prediction of what will happen to you.
It works best for the ancestry groups that dominate research data, and it ignores everything outside your DNA. Treat it as one weak-to-moderate input, not a verdict.
- What a polygenic risk score actually is
- How a score gets built
- Why the number is relative, not absolute
- The ancestry problem
- What a score leaves out
- Polygenic scores versus single-gene headlines
- Are these scores clinically useful yet?
- How to read a score you have been given
- Bottom line
- Common questions
What a polygenic risk score actually is
Some conditions come from one gene. Cystic fibrosis tracks CFTR, and Huntington's disease tracks HTT. Most common conditions are not like that. Heart disease, type 2 diabetes, many cancers and most psychiatric conditions are shaped by a large number of variants that each nudge risk up or down by a small amount, plus diet, sleep, smoking, stress and chance.
A PRS tries to capture the genetic part of that picture by adding up those small nudges. Each person carries several million variants, and only a tiny fraction relate to any given condition. The score is the weighted sum of the ones that do.
The output is usually a percentile or a z-score: "you sit at the 90th percentile of the reference group." That sentence sounds precise. As the rest of this article explains, it hides a lot.
How a score gets built
Researchers run genome-wide association studies, comparing the DNA of many people with a condition against many without it. Variants that show up slightly more often in one group get a weight. A scoring method then combines the weights across the variants it uses, which can number from hundreds to many thousands depending on the method, and the person's own genotype at each position decides how much each one contributes.
Two things follow from that process.
- Nobody needs to know the biology. A score can work statistically without anyone knowing which gene or pathway is responsible. That is a strength for prediction and a weakness for explanation.
- Different tools give different numbers. The variants chosen, the weights and the reference group all vary. Two companies scoring the same person for the same condition can return different percentiles.
Why the number is relative, not absolute
A percentile compares you with a reference population. It does not say how likely you are to develop the condition, or by what age. A 25-year-old and a 75-year-old with the same score have very different real-world risks, because age, sex, blood pressure, cholesterol and family history all carry weight that a score does not capture.
| What a PRS can tell you | What it cannot tell you |
|---|---|
| Where your inherited, common-variant risk sits compared with a reference group | Whether you will actually develop the condition |
| A modest reason to take screening or prevention advice seriously | Your absolute chance over the next ten years |
| A statistical pattern across many people | Why your own body does what it does |
| Nothing about a single rare variant if it is not included | Whether a high-impact gene variant is also present |
The American College of Medical Genetics and Genomics put it plainly in its 2023 points-to-consider statement: a PRS is a statistical estimate of relative risk. A high score does not guarantee disease, and a low score does not rule it out.
The ancestry problem
Most large genetic studies have drawn mainly on people of European ancestry. Scores are tuned on that data, so they tend to be more accurate for people who resemble the study participants and less accurate for everyone else. The same score can mean something different, or very little, for someone whose ancestry is poorly represented.
The NHGRI describes the same limitation, and the ACMG statement says it reduces accuracy for other groups. If your ancestry is mixed or underrepresented, a score may look confident while resting on thin evidence. Ask which population the score was built and tested in. If a report cannot tell you, that is informative.
This is the same equity gap that shows up in uncertain-variant rates, which we cover in our guide to variants of uncertain significance.
What a score leaves out
- Your environment and habits. Diet, activity, smoking, alcohol, medications and sleep often outweigh the inherited common-variant signal.
- Family history. It captures shared genes, shared environment and rare variants that a score may not include. A strong family history can matter more than a middling score.
- Rare, high-impact variants. A PRS summarises common variants. A pathogenic BRCA1 or LDLR variant is a different kind of finding and is not folded into the percentile. The cancer risk guide and our post on familial hypercholesterolemia show how single-gene results behave differently.
- Time. The score does not know how old you are or when risk usually appears.
Do not use a score to rule things out. A "low risk" label is not a reason to skip a screening your doctor recommends. A low percentile only means your common-variant contribution ranks low in that reference group.
Polygenic scores versus single-gene headlines
Popular articles often present one gene as the cause of a condition. The reality for common disease is usually a small effect from each of many variants. Our posts on FTO and obesity and on TCF7L2 and type 2 diabetes show how a famous gene can be statistically real and still explain only a sliver of any one person's outcome. A PRS is the attempt to pool such slivers into something more useful.
APOE sits in between: a small number of variants with a large effect on Alzheimer's risk. It is a single-gene-style result, not a polygenic one, and we explain it in the APOE guide.
Are these scores clinically useful yet?
Mostly not as routine tools. The NHGRI notes that PRS are not yet routinely used by health professionals and that there are no practice guidelines, even though some private and direct-to-consumer companies already sell them. The ACMG statement says evidence that PRS-guided care improves outcomes is limited, advises pretest counselling, and discourages isolated PRS testing when a single-gene cause is known or suspected.
That does not make the idea useless. Research continues, and scores may become more reliable as studies include more ancestries. It does mean you should be wary of a report that turns a percentile into a lifestyle prescription or a scare.
How to read a score you have been given
- Find the condition and the reference group. "Higher than 90% of whom?"
- Check how the score was built and tested, and whether it was validated in people with ancestry like yours.
- Look for absolute numbers. If the report only gives a percentile, you have the least useful half of the story.
- Set it beside what you already know: age, family history, blood pressure, cholesterol, lifestyle.
- Bring it to a clinician or counsellor before changing medication or screening. Our guides to bringing DNA results to your doctor and to genetic counselors cover how.
Bottom line
A polygenic risk score is a legitimate research tool and a shaky personal forecast. It ranks inherited common-variant risk against a reference group, it works unevenly across ancestries, and it says nothing about the rest of your life. If you receive one, read it as a nudge to check the basics that you can actually change, not as a number that decides anything. If you want the broader picture of what a consumer test can and cannot show, start with DTC versus clinical-grade testing.
Common questions
What is a polygenic risk score in simple terms?
It is a number built by adding up the small effects of many common DNA variants to estimate how your inherited risk for a complex condition compares with a reference group. It is a relative ranking, such as a percentile, not a diagnosis or a forecast.
Is a high polygenic risk score a diagnosis?
No. A high score means your inherited common-variant risk ranks high in the reference group. Many people with high scores never develop the condition, and many people with low scores do. Age, family history, health measures and lifestyle all matter.
Why do polygenic risk scores work worse for some ancestries?
Most of the large studies used to build scores have involved people of European ancestry. Scores tuned on that data tend to be more accurate for similar people and less accurate for others, so a result can be misleading if your ancestry is underrepresented.
Should I change my health plan because of a polygenic risk score?
Not on its own. Professional guidance currently says evidence for PRS-guided care is limited. Discuss any result with a clinician or genetic counselor, and keep following the screening and prevention advice that applies to your age and history.
Is a polygenic risk score the same as a single-gene test?
No. A single-gene test looks for a specific high-impact variant, such as one in BRCA1. A polygenic score combines many small-effect variants and does not replace single-gene testing when a specific cause is suspected.
Sources and verification
Facts below were checked in October 2026 against the primary sources listed. The article uses no figures beyond those sources.
Medical disclaimer. GenomeTesting.org publishes educational content about genetics and consumer genomic testing. Nothing here is medical advice, a diagnosis, or a substitute for care from a qualified clinician or genetic counsellor. Decisions about testing, treatment, or interpreting your own results should be made with a healthcare professional who can see your full history.