New AI Tool 'EvORanker' Uses Evolutionary Data to Solve Rare Disease Mysteries

For families of children with rare genetic disorders, the search for answers is often a grueling marathon. Known in the medical community as the "diagnostic odyssey," this journey can stretch for nearly a decade, characterized by endless testing, misdiagnoses, and profound uncertainty. Even with the advent of advanced genetic sequencing, a vast majority of these patients remain undiagnosed, leaving clinicians without a clear path for treatment.[1][3]

Now, a newly developed artificial intelligence tool is poised to radically shorten that timeline. Researchers at the Hebrew University of Jerusalem have introduced EvORanker, an AI algorithm designed to sift through thousands of genetic possibilities and pinpoint the exact mutation responsible for a patient's symptoms. The breakthrough, recently detailed in the journal Genetics in Medicine, represents a major leap forward in how computational biology can be applied to human health.[1][2]

What sets EvORanker apart from conventional diagnostic software is its foundational dataset. Traditional tools rely heavily on existing medical literature and known human genetic databases. If a gene has never been linked to a disease in a published study, standard algorithms will likely miss it. EvORanker, however, looks across the entire spectrum of evolution.[1]

Led by Dr. Christina Canavati and Professor Yuval Tabach, the research team built the AI to analyze how genes have evolved across more than 1,000 different species. The system operates on a well-supported biological principle: genes that evolve together often function together. By detecting these hidden patterns of co-evolution—a technique known as phylogenetic profiling—the algorithm can uncover functional relationships between genes that science has never explicitly linked to human disease before.[1][3][4]

The clinical results have been striking. In rigorous testing, EvORanker identified the correct disease-causing gene as its absolute top candidate in nearly 70 percent of cases. In 95 percent of the cases, the culprit gene was placed within the top five candidates. This performance significantly outpaces existing diagnostic tools, particularly in complex scenarios involving poorly understood or entirely novel genetic variants.[1][4]

Beyond the impressive statistics, the tool is already changing lives. The study highlighted a case involving a child with a complex, undiagnosed neurodevelopmental disorder. After conventional diagnostics had completely failed, researchers ran the patient's data through EvORanker. The AI identified a previously unrecognized gene as the likely cause, finally providing the family with an answer and giving doctors a target for potential therapeutic strategies.[1]

In another instance, the algorithm successfully revealed the genetic basis of a severe multisystem disease affecting multiple organs. By identifying biallelic variants in the SUPT4H1 gene, the tool helped clinicians understand that the disorder was tied to disrupted transcription processes. This level of clarity is crucial, as identifying the exact mechanism of a disease is the first necessary step toward developing a targeted treatment.[2][5]

The implications of this technology extend far beyond initial diagnoses. By uncovering new disease-causing genes, EvORanker can help researchers identify existing, FDA-approved drugs that might be repurposed to treat these rare conditions. This shortcut could save years of pharmaceutical development time, bringing effective therapies to patients much faster than the traditional drug discovery pipeline allows.[1][3]

While rare diseases are the immediate focus—affecting up to 5 percent of the global population and up to 8 percent in certain communities in Israel—the research team is already looking ahead. They are currently exploring how the EvORanker technology can be applied to oncology. Specifically, researchers hope to use the algorithm to understand why some tumors unexpectedly regress, potentially unlocking new mechanisms for treating stage 4 cancer.[1][4]

EvORanker is now available as an accessible tool for researchers and clinicians worldwide. As more medical institutions integrate evolutionary AI into their diagnostic workflows, the agonizing years of uncertainty that define the rare disease experience may soon become a relic of the past.[1][4]