Prion disease is rare, and its studies are small. The completed drug trials enrolled between 7 and 121 people each, and the active PrP-lowering trials are early. Any model trained or evaluated on this field works with limited data, and the Atlas shows the stage and source of each claim so that it can be cited accurately. See problems worth solving and data availability.
Tools that answer questions about prion disease are used by frightened families and by clinicians. They should state facts with sources, avoid interpreting any person’s symptoms, tests or genetic results, and send people to a specialist or genetic counsellor. See responsible use and machine access. Wider context is on the partners page, with funders and policy alongside.
Explore
- Data availabilityWhat prion disease data can actually be obtained, on what terms, and which datasets a model developer will not find.
- Problems worth solvingWhere AI and data science could help in prion disease: small datasets, late diagnosis and endpoints, and what is not yet known.
- Responsible use of AI in this fieldExpectations for AI tools used with prion disease information: cite sources, avoid interpreting individual results, and refer people to specialists.