Polygenic risk scores explained

What a polygenic score is

Most common conditions aren’t caused by a single gene — they’re influenced by the combined, individually small effect of many genetic variants together. A polygenic score adds up those effects, weighted by how strongly each variant is associated with the condition in large genetic studies, into a single number. On its own that number means little; what matters is where it falls compared to other people — a percentile.

The 18 conditions DNA Layer scores

Each condition maps to one of DNA Layer’s health areas:

• Heart health: Coronary Artery Disease, Atrial Fibrillation, Ischemic Stroke, Venous Thromboembolism
• Metabolic health: Type 2 Diabetes, Gout
• Diseases & prevention: Breast Cancer, Prostate Cancer, Colorectal Cancer, Melanoma, Macular Degeneration, Chronic Kidney Disease
• Mood & brain: Alzheimer's Disease, Parkinson's Disease, Major Depression, Migraine
• Immune system & inflammation: Asthma
• Weight loss: Obesity

How the percentile is computed

The weights come from published GWAS-derived scoring files (PGS Catalog). Your genotypes are scored against those weights, then ranked as a percentile against a reference population — scored on exactly the variants your file covers, so partial coverage doesn’t bias the comparison. The percentile carries a confidence interval, not a single false-precise number. These scores are attached to your report and its agentic export where a condition has enough covered variants to score reliably.

What this number is not

The same three limits appear on every polygenic card DNA Layer shows, at every surface — not just here:

• This is not a diagnosis and it cannot tell you whether you will develop this condition.
• It measures common variants only — it does not look for rare high-impact mutations.
• Lifestyle, environment, age, and family history are not part of this number.

One limitation worth naming explicitly, since it’s common across the polygenic score field and not always disclosed: the comparison group is of European ancestry, so the ranking is less reliable for people of other genetic ancestries. This is a known, documented limitation of the underlying reference data available today, not something specific to any one condition.