The Evidence Aggregator: AI reasoning applied to rare disease diagnostics.
An AI tool rapidly extracts disease-variant evidence from research literature, potentially speeding diagnosis for people with rare genetic diseases.
EvAgg, an open-source generative-AI tool from Microsoft Research and the Broad Institute, systematically extracts variant-disease evidence from scientific literature and demonstrated 92% paper recall, 96% variant detection recall, and 34% time savings in a user study with rare disease geneticists. This near-term deployable tool addresses the bottleneck of manual literature review in rare disease diagnostics, potentially reducing diagnostic latency and increasing diagnostic solve rates for challenging cases.
What the study was
- Study design
- AI tool validation study with expert-curated dataset and user study
- Population
- Rare disease genetics cases; expert clinical geneticists as users
- Category
- Diagnostics
- Maturity
- Validated
- Journal
- Genet Med
Why it surfaced
Open-source tool from Microsoft/Broad already implemented; user study demonstrates 34% time savings in rare disease case analysis. Genet Med is the flagship ACMG journal. Tool spans AI/ML diagnostics and rare disease topics.
A plain-language summary of published research — not medical advice. Talk to a clinician about your own care.