Are the data reliable?
Sample and variant QC, harmonisation, and imputation checks establish whether genotype data are suitable for the selected model.
Science & quality
A useful polygenic score needs more than a strong association: reliable data, evidence for the intended population, and clear limits on interpretation.
01 · Understand the score
Inherited predisposition. In context.
A PRS estimates inherited predisposition relative to a reference population. It is not a diagnosis or a prediction of an individual outcome. Interpretation depends on calibration, ancestry, and intended use.
02 · Evaluate the evidence
Three quality checks.
Does the model fit the population?
Discrimination, calibration, independent validation, and ancestry transferability are assessed for the intended population and use. A published association alone is not enough.
Can the result be traced?
Model specifications, reference data, pipeline versions, and quality criteria remain linked to the result so its basis can be reviewed.
03 · Apply with clinical context
Clinical context comes first.
Interpret PRS alongside age, family history, biomarkers, and clinical measurements. Its value depends on what it adds to the existing care pathway; limitations and validation boundaries must remain explicit.
Discuss your intended use