Once large GWAS studies identified hundreds of schizophrenia-associated variants, the next obvious question was whether those variants could be combined into a single number predicting risk. That number is the polygenic risk score (PRS). Over the past decade, PRS has become one of the most discussed — and most misunderstood — tools in psychiatric genetics.
A polygenic risk score for schizophrenia sums the small effects of many common genetic variants weighted by their GWAS-derived contributions, producing a single number that explains some but not most of the variation in risk between individuals.
How a PRS is calculated
A polygenic risk score is built in two steps. First, a large GWAS — typically from the Psychiatric Genomics Consortium — produces effect-size estimates for millions of SNPs. Second, in a separate person, each of those SNP genotypes is multiplied by its weight and summed. The result is a single number that places the person somewhere on a risk distribution. People in the top decile of schizophrenia PRS have several times the average population risk; people at the bottom have less.
What the research shows
Several large studies have characterised what schizophrenia PRS can and cannot do:
- Population-level discrimination is modest. Schizophrenia PRS, in current models, explains roughly 7% to 12% of variance in schizophrenia risk on the liability scale. The area under the receiver operating characteristic curve (AUC) for distinguishing cases from controls is around 0.7 — better than chance, well below clinical diagnostic standards.
- Risk stratification works. People in the highest PRS decile have several-fold higher risk than those in the lowest. This holds in datasets like the UK Biobank.
- PRS is enriched in first-episode and treatment-resistant cases. Patients with schizophrenia tend to have higher PRS than healthy controls; some studies suggest higher scores in those with poorer treatment response, although findings vary.
- PRS shows cross-disorder overlap. Schizophrenia PRS is also elevated in people with bipolar disorder, severe depression, and to a lesser extent autism.
The 2022 PGC paper in Nature remains the standard reference for current discriminative performance.
Why PRS is not yet clinical
Several barriers separate research-grade PRS from a clinical screening tool:
- Insufficient predictive power. A test that correctly classifies people with around 70% accuracy is far below the bar for psychiatric screening, where the consequences of false positives include unnecessary anxiety and stigma.
- Population specificity. Most schizophrenia GWAS data has come from European-ancestry samples. PRS calibrated on those data performs notably worse in African, East Asian, South Asian, and other ancestries — a problem documented in studies like the 2019 Nature Genetics review by Martin and colleagues.
- No actionable intervention. Even a perfectly accurate PRS would only matter if knowing the score changed care. Currently, no preventive intervention is approved for high-PRS individuals who do not have symptoms.
- Ethical concerns. Predicting a stigmatised condition like schizophrenia in an asymptomatic person — especially a child — raises questions about consent, insurance, employment, and self-image.
Where PRS is genuinely useful
Despite clinical limitations, PRS has real value in research:
- Stratifying clinical high-risk samples. In studies of people with prodromal symptoms, PRS may help identify who is at highest risk of conversion to a full psychotic disorder, in combination with clinical features.
- Studying drug response and side effects. Researchers are investigating whether PRS predicts response to clozapine, weight gain on certain antipsychotics, or risk of cardiovascular complications.
- Distinguishing diagnoses. PRS can support cross-disorder research by quantifying genetic overlap between conditions.
- Mechanism research. Individuals at the extremes of PRS distributions are useful for studying the biology of risk.
Direct-to-consumer offerings
Some commercial companies now offer polygenic scores for psychiatric conditions including schizophrenia. The American Psychiatric Association, the National Society of Genetic Counselors, and most academic psychiatric geneticists have urged caution. The scores often lack rigorous validation, are based on European-ancestry data, and risk causing distress without offering an evidence-based action.
Talk to a board-certified genetic counsellor. Ask what the test is validated to do, what proportion of risk it explains, and what — if anything — you would change based on the result.
How PRS compares to family history
For most people, a careful family history remains a more useful guide to schizophrenia risk than current PRS. Having a first-degree relative with schizophrenia raises lifetime risk to roughly 10% (compared with about 1% in the general population). PRS captures complementary information — not all genetic risk is shared with relatives — but neither tool is precise enough to be deterministic. See our piece on genetic risk of schizophrenia for the family-history perspective.
The bottom line
Polygenic risk scores for schizophrenia are a powerful research tool and an immature clinical one. They will likely improve as sample sizes grow, ancestry diversity widens, and methods refine. They are unlikely to ever be a single deterministic predictor — schizophrenia is too multifactorial. The most realistic future is one in which PRS is one input among many — alongside family history, clinical features, environmental exposures, and possibly biomarkers — that guides personalised prevention and treatment for those at highest risk.
This article is for educational purposes only and is not medical advice, diagnosis, or treatment. Always consult a qualified mental health professional. If you or someone you know is in crisis, call or text 988 in the US, or your local emergency number.