Polygenic Risk Scores Explained: Beyond Single-Gene Tests
How thousands of tiny genetic effects combine to shape your disease risk — and what this means for you
By GenomeInsight Science Team
Key Takeaways
- PRS combines the effects of thousands of small-effect genetic variants into a single risk score — unlike single-gene tests that look for rare high-impact mutations
- Top 5% PRS for coronary artery disease confers risk equivalent to familial hypercholesterolemia (~3× lifetime risk)
- Ancestry bias is the biggest limitation — most PRS are derived from European-ancestry studies and perform worse in other populations
- PRS is not diagnostic — it shifts probabilities but doesn't determine outcomes; lifestyle still matters enormously
- Clinical use is most advanced for coronary artery disease, breast cancer, type 2 diabetes, and prostate cancer
- A favorable lifestyle can offset ~46% of high genetic risk for heart disease — genes are not destiny
From Single Genes to Polygenic Risk
Traditional genetic testing focuses on single-gene (Mendelian) conditions — like BRCA1 for breast cancer, or the F5 gene for Factor V Leiden. These are powerful: a single pathogenic variant dramatically changes your risk. But they're also rare. Only about 5-10% of most common diseases are explained by high-impact single-gene mutations.
For the other 90-95%, the genetic architecture looks completely different. Conditions like heart disease, type 2 diabetes, Alzheimer's, depression, and most cancers are polygenic — influenced by hundreds or thousands of genetic variants, each contributing a tiny amount of risk.
Think of it this way:
- •Mendelian risk: One boulder rolls onto the scale and tips it. Rare but dramatic.
- •Polygenic risk: Thousands of grains of sand accumulate. None is individually significant, but together they shift the balance meaningfully.
A Polygenic Risk Score (PRS) captures this second type of risk. It sums up the effects of thousands (sometimes millions) of common genetic variants to produce a single number representing your genetically predicted risk for a particular condition.
How Polygenic Risk Scores Work
Building a PRS involves several steps:
Step 1: Genome-Wide Association Studies (GWAS) Scientists scan the genomes of hundreds of thousands of people — some with a disease, some without — to identify which SNPs (single nucleotide polymorphisms) are statistically associated with the condition. Modern GWAS studies include 500,000+ participants and identify hundreds to thousands of associated variants.
Step 2: Effect Size Estimation For each associated SNP, the GWAS estimates an effect size (odds ratio or beta coefficient) — how much that variant shifts risk. Most individual effects are tiny: an odds ratio of 1.01-1.10 (1-10% risk change per variant).
Step 3: Score Calculation Your PRS is calculated by counting how many risk-increasing alleles you carry at each associated position and weighting each by its effect size:
PRS = Σ (effect_size_i × allele_count_i) for all variants i = 1 to n
Where n can range from a few hundred to several million variants, depending on the method.
Step 4: Population Percentile Your raw PRS is converted into a percentile by comparing it to a reference population. If you're in the 90th percentile for coronary artery disease PRS, your genetically predicted risk is higher than 90% of the reference population.
What the percentile means in practice:
| PRS Percentile | Typical Risk Interpretation |
|---|---|
| <20th | Lower than average genetic risk |
| 20th-80th | Average genetic risk |
| 80th-95th | Moderately elevated genetic risk |
| >95th | Substantially elevated genetic risk (~2-3× for many conditions) |
| >99th | Highly elevated genetic risk (can approach monogenic risk levels) |
Current Clinical Use of PRS
As of 2025-2026, polygenic risk scores are transitioning from research tools to clinical applications. Here's where things stand:
Coronary Artery Disease (CAD): The most clinically advanced PRS. Individuals in the top 5% of PRS have risk equivalent to having familial hypercholesterolemia (~3× lifetime risk). The AHA/ACC guidelines now acknowledge PRS as a "risk-enhancing factor" that can inform statin therapy decisions. Several health systems (including the UK NHS and Geisinger Health) are piloting CAD PRS in primary care.
Breast Cancer: PRS is being integrated into breast cancer screening algorithms. Women in the top 1% of breast cancer PRS have ~3.5× risk — comparable to having a first-degree relative with breast cancer. The BOADICEA model (used by the UK NHS) now incorporates PRS alongside traditional risk factors and BRCA status to personalize mammography screening schedules.
Type 2 Diabetes: T2D PRS can identify high-risk individuals years before disease onset. When combined with clinical risk factors (BMI, family history, fasting glucose), PRS significantly improves prediction. Those in the top decile have ~3× risk even with normal BMI — challenging the assumption that T2D is purely lifestyle-driven.
Prostate Cancer: PRS is being integrated into prostate cancer screening decisions, helping determine who benefits from PSA screening and at what age to start. The top 2% of PRS have ~5× lifetime risk.
Other areas in development: Atrial fibrillation, inflammatory bowel disease, Alzheimer's disease, schizophrenia, and many more conditions have validated PRS models in various stages of clinical translation.
Important Limitations
PRS is powerful, but it has significant limitations that must be understood:
1. Ancestry Bias — The Biggest Problem
- •A CAD PRS might explain ~15% of risk variance in Europeans but only ~4% in African Americans
- •Effect sizes at specific loci may differ across populations
- •Linkage disequilibrium patterns vary, meaning the "tag" SNPs identified in Europeans may not capture the same signals in other ancestries
This is an active area of research. Multi-ancestry GWAS and transfer learning methods are improving cross-ancestry PRS performance, but the gap remains significant.
2. Missing Heritability
- •Height PRS explains ~40% of variance (the best-performing PRS)
- •CAD PRS explains ~15-20% of variance
- •Most disease PRS explain 5-15% of variance
The remaining genetic contribution comes from rare variants, gene-gene interactions, gene-environment interactions, and structural variants not captured by GWAS arrays.
3. PRS Is Not Diagnostic A high PRS does not mean you WILL develop the condition. A low PRS does not mean you WON'T. PRS shifts probabilities — it doesn't determine outcomes. Lifestyle, environment, and other genetic factors all matter.
4. Limited Actionability for Some Conditions For conditions where PRS is validated (CAD, breast cancer), clinical actions are clear: screening, lifestyle modification, or preventive medication. For others (schizophrenia, depression), high PRS is informative but actionable options are less defined.
5. Environment and Lifestyle Still Matter A high genetic risk score for coronary artery disease can be substantially offset by a healthy lifestyle. The Khera et al. (2016) study showed that among those in the top quintile of genetic risk for CAD, a favorable lifestyle reduced risk by ~46%.
What GenomeInsight's PRS Covers
GenomeInsight calculates polygenic risk scores from your consumer genotyping data for conditions where PRS has robust scientific validation:
Currently included:
- •Coronary Artery Disease — based on the Khera et al. genome-wide PRS (millions of variants)
- •Type 2 Diabetes — multi-variant score validated across populations
- •Breast Cancer — 313-variant PRS validated in large cohorts
- •Prostate Cancer — based on the latest PRACTICAL consortium data
- •Atrial Fibrillation — clinically relevant for stroke risk assessment
- •Alzheimer's Disease — incorporating both APOE and non-APOE polygenic risk
How we present results:
- •Your population percentile (compared to the appropriate ancestry reference)
- •Absolute risk estimate (not just relative — we show your estimated lifetime risk)
- •Ancestry context — we note if the PRS has limited validation for your ancestral background
- •Actionable recommendations — what the score means for screening, lifestyle, and medical follow-up
- •Confidence level — we distinguish between well-validated scores and those still in the research phase
What we DON'T do:
- •We don't present PRS for conditions without sufficient validation
- •We don't use PRS from a single ancestry to make claims about other ancestries without appropriate caveats
- •We don't present PRS in isolation — they're shown alongside single-gene results, family history, and lifestyle factors for a complete picture
The Future of PRS in Medicine
Polygenic risk scores are evolving rapidly. Here's what's coming:
Near-term (2025-2027):
- •Clinical integration of CAD and breast cancer PRS into standard primary care workflows
- •Multi-ancestry PRS becoming standard (moving beyond European-only scores)
- •Newborn/childhood PRS screening pilots (e.g., the eMERGE network and BabySeq studies)
- •Pharmacogenomics + PRS combination (predicting drug response using both approaches)
Medium-term (2027-2030):
- •PRS incorporated into standard health insurance risk assessments (with regulatory protections needed)
- •Dynamic PRS that integrate epigenetic modifications and biomarker data
- •Population-scale PRS screening programs (following the UK Biobank and All of Us models)
- •PRS-guided clinical trials (enrolling based on genetic risk)
Longer-term (2030+):
- •Whole-genome sequencing replacing array-based PRS (capturing rare variants)
- •Gene-environment interaction scores (combining PRS with wearable/lifestyle data)
- •Truly global PRS validated across all ancestries
- •Integration with other -omics data (proteomics, metabolomics) for comprehensive risk models
Ethical considerations:
- •Insurance discrimination based on genetic risk (GINA covers health insurance but not life/disability)
- •Equity of access — PRS must work for all populations, not just Europeans
- •Psychological impact of probabilistic risk information
- •Privacy of genetic data in an era of large biobanks
GenomeInsight is committed to advancing PRS responsibly — providing accurate, contextualized, and equitable genetic risk information to everyone.
Medical Disclaimer: This article is for educational purposes only and does not constitute medical advice. Genetic information should be interpreted in the context of your full medical history by a qualified healthcare provider. Never change medications without consulting your doctor.
References
- [1]Khera AV et al. (2018). Genome-wide polygenic scores for common diseases identify individuals with risk equivalent to monogenic mutations. Nat Genet. 50(9):1219-1224.PubMed
- [2]Martin AR et al. (2019). Clinical use of current polygenic risk scores may exacerbate health disparities. Nat Genet. 51(4):584-591.PubMed
- [3]Khera AV et al. (2016). Genetic Risk, Adherence to a Healthy Lifestyle, and Coronary Disease. N Engl J Med. 375(24):2349-2358.PubMed
- [4]Torkamani A et al. (2018). The personal and clinical utility of polygenic risk scores. Nat Rev Genet. 19(9):581-590.PubMed
Check your genetic variants
Upload your 23andMe or AncestryDNA raw data to see your personal results. Processing happens entirely in your browser — no data leaves your device.
Analyze Your DNA Free