Polygenic risk scores
A PRS combines many common genetic variants into a single model-based score. In this report, the percentile tells you where the sample falls relative to the reference population used by each study model.
Important Notice
The results shown on this website are provided for research and educational purposes only. They are not a clinical diagnosis, they do not replace medical advice, and they must not be used to make healthcare decisions.
Prostate Cancer PRS
This client sends a .vcf or .vcf.gz file to the backend, starts a
job, tracks progress, and renders the PRS results grouped by study.
Research-oriented genomic interpretation.
A PRS combines many common genetic variants into a single model-based score. In this report, the percentile tells you where the sample falls relative to the reference population used by each study model.
Different studies can disagree because they use different populations, variant sets, and statistical weights. This interface shows that disagreement instead of hiding it.
When available, the report also adds monogenic panel findings and AlphaGenome-based functional notes to help explain which signals may deserve closer attention in research.
Step-by-step overview of the scoring workflow.
Watch the animation...
Short guide to the main outputs.
A higher percentile means the sample ranks higher within the reference distribution used by that specific PRS model. It is a comparative position, not a diagnosis.
The study-level z-score reflects how far the sample's PRS sits from the model's reference average. Larger absolute values mean a more extreme score within that study.
Variant-level z-scores indicate which matched variants contributed most strongly to the final PRS in a given study. They are contribution measures, not p-values.
AlphaGenome outputs are model-based functional predictions. They can help prioritize biologically interesting signals, but they are not clinical proof on their own.
What happens after upload.
Core score models currently included in the platform.
Questions, collaboration, or feedback.
If you want to discuss methodology, report wording, or possible improvements to the platform, you can get in touch directly.
Research collaboration, bug reports, feature requests, and clarification about generated reports.