Hebrew University AI tool EvORanker helps pinpoint the genes behind rare diseases

Researchers at the Hebrew University of Jerusalem have released an artificial intelligence tool called EvORanker that helps doctors find which gene is responsible for a patient's rare disease. The work, led by Dr. Christina Canavati and Prof. Yuval Tabach of the university's Faculty of Medicine, was described in the journal Genetics in Medicine and announced by the university at the end of March 2026.

When a child has an unexplained genetic condition, sequencing usually turns up thousands of possible variants. Most diagnostic software ranks them using what is already known about each gene, so genes that have never been linked to illness tend to be overlooked. EvORanker takes a different route. It looks at how genes have been gained and lost together across more than 1,000 species, on the idea that genes which evolve in step usually work together. That lets it flag a likely culprit even when no human case has been recorded before.

In clinical testing reported by the team, the correct gene came out as the top candidate in nearly 70% of cases and appeared in the top five in about 95%. The advantage over existing tools was largest for poorly studied genes. In one case, a child with a complex neurodevelopmental disorder had been through extensive testing without an answer, and the tool pointed to a gene not previously tied to disease. In another, it uncovered the genetic basis of a severe disorder affecting several organs and suggested possible treatment strategies.

The method builds on evolutionary gene-comparison research that Tabach carried out with Nobel laureate Gary Ruvkun, published in 2013. Rare diseases together affect up to about 5% of people worldwide, and families often wait close to a decade for a diagnosis. EvORanker is now open to clinicians and researchers, and the group says it is also exploring uses in cancer research and in finding existing drugs that could be repurposed.