Polygenic Risk Scores Explained: What They Can and Can't Tell You

TL;DR: A polygenic risk score (PRS) aggregates the effects of thousands of genetic variants into a single number estimating your predisposition to a disease or trait. PRS can identify individuals at several-fold increased risk for conditions like heart disease and breast cancer, but they are probability estimates — not diagnoses — and their accuracy varies significantly across ancestral populations.

Disclaimer: This article is for educational purposes. It does not constitute medical advice. Consult a healthcare professional for personalized guidance.

When people think of genetic risk, they often picture single dramatic mutations — a BRCA1 variant that sharply increases breast cancer risk, or the Huntington's disease gene that all but guarantees onset. But for most common diseases, the genetic architecture looks nothing like that. Heart disease, type 2 diabetes, and many cancers are influenced by thousands of genetic variants, each shifting risk by a tiny amount. Polygenic risk scores were developed to capture this distributed signal. They compress the contributions of hundreds of thousands of SNPs into a single number that estimates where you fall on the spectrum of genetic risk for a given condition. As part of a broader DNA analysis, PRS offers a way to quantify what was previously unmeasurable.

What Is a Polygenic Risk Score?

A polygenic risk score is a single numerical value that estimates an individual's genetic predisposition to a specific trait or disease based on the cumulative effect of many genetic variants across the genome. Unlike single-gene tests that look for one high-impact mutation, a PRS weighs the small contributions of thousands — sometimes millions — of common SNPs identified through genome-wide association studies.

Polygenic risk score (PRS): A quantitative estimate of an individual's genetic liability for a trait or disease, calculated by summing the effects of many genetic variants, each weighted by its effect size from genome-wide association studies.

Think of it like a credit score for genetic risk. No single factor — no single SNP — determines the overall score. But the aggregate pattern across thousands of data points produces a meaningful signal. A high PRS for coronary artery disease does not mean you will have a heart attack. It means that, on average, people with your genetic profile develop heart disease more often than people with lower scores.

The key distinction is between PRS and monogenic testing. A pathogenic BRCA1 variant confers a lifetime breast cancer risk of 60-80%. That is a strong, actionable signal from a single gene. A high PRS for breast cancer, by contrast, reflects the combined push of many variants, each contributing a fraction of a percent. The aggregate effect can be substantial — the top 1% of the breast cancer PRS distribution faces up to 4-fold higher risk than average — but the mechanism is fundamentally different: distributed rather than concentrated.

How Are Polygenic Risk Scores Calculated?

The mathematics behind PRS is conceptually straightforward, even if the implementation requires careful statistical work.

Step 1: Researchers conduct a genome-wide association study (GWAS) on a large cohort — often hundreds of thousands of people — comparing the frequency of millions of SNPs between individuals with and without the trait of interest. Each SNP receives an estimated effect size (beta coefficient) reflecting its association strength.

Step 2: For a target individual, their genotype at each SNP position is multiplied by that SNP's effect size from the GWAS.

Step 3: These weighted values are summed across all included SNPs to produce a single continuous score.

In notation: PRS = the sum of (genotype at SNP-i multiplied by beta-i) across all i SNPs. A person's raw PRS is then typically compared to a reference distribution to express it as a percentile.

Genome-wide association study (GWAS): A study design that tests millions of genetic variants across large populations to identify which are statistically associated with a disease or trait, producing effect size estimates used in polygenic scoring.

The devil is in the details. Early PRS methods used a simple "clump and threshold" approach: select the most strongly associated SNPs after removing redundant ones correlated through linkage disequilibrium, then sum their effects. Modern methods like LDpred and PRS-CS use Bayesian frameworks to jointly model all SNPs while accounting for their correlations, generally producing more accurate scores. The PGS Catalog, a public repository launched in 2021, now hosts thousands of validated polygenic scores across hundreds of traits.

What Can Polygenic Risk Scores Tell You?

Coronary Artery Disease

The strongest clinical evidence for PRS exists in cardiovascular disease. A landmark 2018 study by Khera and colleagues, published in Nature Genetics, found that 8% of the population falls in a PRS range conferring more than 3-fold increased risk for coronary artery disease — a risk level comparable to that of carriers of rare familial hypercholesterolemia mutations. Notably, this high-risk PRS group is 20-fold more prevalent than the monogenic mutation carriers, meaning PRS identifies far more at-risk individuals than traditional genetic testing.

More recent work has refined this. A 2025 study in npj Cardiovascular Health showed that integrating PRS with conventional clinical risk scores reclassified 7-10.7% of patients previously categorized as borderline or intermediate risk into a high-risk group, with hazard ratios of 3.20 to 3.84 for coronary events. For people in this clinical gray zone, PRS adds information that standard risk calculators miss.

Breast Cancer

Polygenic risk scores for breast cancer can stratify women beyond what family history alone provides. Women in the top 1% of the PRS distribution face up to a 4-fold increase in risk compared to those at average risk. Clinical frameworks now exist to combine PRS with other factors — including mammographic density and family history — to personalize screening intervals and starting ages.

Type 2 Diabetes

PRS for type 2 diabetes identifies individuals at elevated genetic risk before any clinical signs appear. Combined with lifestyle and metabolic markers, this information can target preventive interventions — diet modifications, exercise programs, earlier glucose monitoring — to those most likely to benefit. The Khera et al. study found that 1.5% to 8% of the population carries PRS-defined risk equivalent to monogenic forms of each studied disease, including type 2 diabetes.

What Polygenic Risk Scores Cannot Tell You

They Are Not Diagnoses

A PRS is a probability estimate, not a crystal ball. A person in the top 5% of the coronary artery disease PRS distribution has a meaningfully higher genetic risk — but if that person exercises regularly, maintains a healthy weight, does not smoke, and manages their cholesterol, their absolute risk of a cardiac event may be lower than someone with an average PRS who does none of those things.

Environment, behavior, and lifestyle interact with genetic predisposition in ways PRS cannot capture. The score tells you about one input to a complex equation. Treating it as a verdict is a misuse of the tool.

The Ancestry Gap

This is the most serious limitation facing PRS today. Between 2008 and 2017, 67% of polygenic scoring studies included exclusively European-ancestry participants. Another 19% focused on East Asian ancestry. Only 3.8% included individuals of African, Hispanic, or Indigenous descent, according to a 2019 analysis in Nature Communications.

The consequence is predictable: PRS accuracy declines as an individual's genetic ancestry diverges from the European-dominant training data. A 2023 study in Nature demonstrated that polygenic scoring accuracy decreases along the continuum of genetic ancestries, with the steepest drops for individuals of African descent. Deploying European-trained PRS in diverse clinical settings without adjustment risks exacerbating existing health disparities — the opposite of precision medicine's promise.

The field is actively working to close this gap. The All of Us Research Program, the Africa Wits-INDEPTH Partnership for Genomic Studies, and expanding multi-ancestry biobanks are generating more representative datasets. Multi-ancestry PRS methods that integrate GWAS data from multiple populations show improved transferability, but achieving equitable performance remains years away.

Limited Individual-Level Precision

PRS performs better at the population level than the individual level. It can reliably identify groups of people at higher average risk, but the overlap in score distributions between people who do and do not develop disease is substantial. A 2023 appraisal in the Journal of Internal Medicine found that the diagnostic and prognostic performance of PRS alone is consistently low. Combining PRS with conventional clinical risk factors produces moderate improvement — meaningful, but not transformative.

PRS in Clinical Practice Today

Despite these limitations, polygenic risk scores are entering clinical workflows. The eMERGE Network — a consortium funded by the National Human Genome Research Institute — has developed and validated PRS for ten chronic diseases across 25,000 diverse adults in the United States. The selected conditions include coronary artery disease, type 2 diabetes, atrial fibrillation, and breast cancer, chosen based on PRS performance and clinical actionability.

PRS has several practical advantages as a clinical tool: it is inexpensive (derivable from a standard SNP array), available from birth, and only needs to be measured once in a lifetime. The main barriers to widespread adoption are not technical but operational — clinician education, patient communication, integration into electronic health records, and establishing clear guidelines for when and how to act on PRS results.

Cardiovascular medicine is furthest along. The European Society of Cardiology and several US institutions have begun incorporating PRS into risk stratification for patients with borderline clinical risk profiles. For pharmacogenomics — where the question is drug response rather than disease risk — the evidence base for SNP-guided clinical decisions is already more mature.

At DeepDNA, we view PRS as one component of a broader genetic risk assessment. The score is most valuable when contextualized alongside single-gene findings, pharmacogenomic variants, and modifiable lifestyle factors. A PRS in isolation is an incomplete picture; combined with actionable insights about drug metabolism and nutrient processing, it becomes part of a practical health strategy.

Frequently Asked Questions About Polygenic Risk Scores

Can a polygenic risk score predict if I will get a disease?
No. A PRS estimates your relative genetic predisposition compared to a reference population. It does not predict individual outcomes. Two people with identical PRS values can have very different health trajectories based on lifestyle, environment, and chance.

How is a PRS different from genetic testing for BRCA?
BRCA testing looks for rare, high-impact mutations in a single gene. A PRS aggregates the small effects of thousands of common variants across the genome. Both provide risk information, but through fundamentally different genetic mechanisms — concentrated effect versus distributed effect.

Are polygenic risk scores available to consumers?
Yes. Several direct-to-consumer genetic testing services now include PRS reports for selected conditions. However, the clinical utility of consumer PRS varies by provider, and results should be discussed with a healthcare professional for proper interpretation.

Will my PRS change over time?
Your genetic variants do not change, so your raw score remains fixed. However, the interpretation and accuracy of PRS will improve as researchers analyze larger, more diverse datasets and refine statistical methods. A score calculated today may be reinterpreted with greater precision in the future.

The Future of Genetic Risk Assessment

Polygenic risk scores represent a genuine advance in genetic risk assessment — and a genuinely imperfect one. They can identify substantial proportions of the population at elevated risk for major diseases, risk levels sometimes comparable to rare monogenic mutations that clinicians already act on. But they are probability estimates that work best at the population level, perform unevenly across ancestries, and capture only the genetic slice of a multifactorial equation.

The trajectory is toward improvement. Larger and more diverse GWAS, better statistical methods, and integration with clinical and lifestyle data will make PRS more accurate and more equitable. The challenge for the field is to deploy these tools responsibly — communicating uncertainty clearly, ensuring equitable access, and never confusing a risk score with a diagnosis.

DeepDNA integrates polygenic risk insights with SNP-level analysis, pharmacogenomic profiling, and ancestry data — all within a framework designed for European data privacy standards. Understanding your genetic risk is the starting point; knowing what to do with that information is where the value lies.

---

Originally published at deepdna.ai