Close Menu
    2digital.news2digital.news
    • News
    • Analytics
    • Interviews
    • About us
    • Editorial board
    • Events
    2digital.news2digital.news
    Home»News»Artificial intelligence “learns from the tree of life.” A new model, popEVE, helps diagnose rare diseases using AI
    News

    Artificial intelligence “learns from the tree of life.” A new model, popEVE, helps diagnose rare diseases using AI

    Mikolaj LaszkiewiczBy Mikolaj LaszkiewiczNovember 24, 20253 Mins Read
    LinkedIn Twitter Threads Reddit
    Share
    Twitter LinkedIn Threads Reddit

    The rapid development of AI brings risks, but it can also deliver major benefits. Researchers from the Centre for Genomic Regulation in Barcelona and Harvard Medical School used an AI model called popEVE for exactly that purpose — to identify mutations potentially responsible for rare diseases.

    The team built popEVE on the earlier EVE algorithm from 2021, but the new version significantly expands diagnostic capabilities. Unlike conventional approaches, which require population data and case descriptions, popEVE is based on analyzing evolutionary conservation of proteins. It uses massive sets of amino-acid sequences from tens of thousands of species to determine which positions in proteins have remained unchanged for millions of years. When a mutation appears in a highly conserved site, the AI assesses it as potentially pathogenic.

    This approach is groundbreaking for diagnosing rare diseases, where many mutations fall into the category of VUS (“variant of uncertain significance”). In clinical testing, popEVE analyzed data from 513 children with severe neurodevelopmental disorders and correctly identified the most likely pathogenic mutation in 98% of cases. This is an excellent result for a tool that operates without family data and without large comparative cohorts.

    As Alexander Morozov (physician, researcher, and expert in the convergence of technology and medicine) told us: “This work describes a development that has the potential to significantly change how rare diseases are diagnosed. The ability to assess the pathogenicity of mutations based on evolutionary context is a strong step forward. Since population databases and clinical panels still don’t cover the full diversity of human mutations, a tool capable of handling ‘unknown’ variants looks extremely promising.

    If the technology truly scales and proves itself in clinical practice, it could remove one of the major barriers — the absence of parental samples and the long ‘journey’ a patient often takes between specialists in search of a diagnosis. For rare diseases, speed matters immensely: the earlier the diagnosis is made, the higher the chance of initiating appropriate therapy on time or at least planning proper monitoring.

    Of course, what we’re seeing now is only an early stage. More data is needed on how reliably the system performs in real clinical settings, how it handles complex or mixed genetic cases, and how consistent the results are across different centers. Even so, the direction of progress is truly inspiring — tools like this could dramatically reduce the diagnostic odyssey for patients.”

    Importantly, the model also performs well in situations where a mutation is unique to a single patient — something that poses a major challenge for traditional diagnostic methods. The researchers emphasize that popEVE can also be used in countries with limited technological resources, because its effectiveness relies on evolutionary data that is globally accessible and independent of local infrastructure. It does not require large, expensive data centers to operate.

    The impact of the project goes far beyond diagnostics alone. Rare diseases affect up to 300 million people worldwide, and more than half of patients never receive a definitive diagnosis. Much of this is due to the lack of sufficient clinical data to evaluate rare mutations. PopEVE fills this gap by offering a mechanism grounded in universal principles of protein evolution rather than in the availability of medical datasets.

    The authors stress that in the future, popEVE could be integrated into routine genetic diagnostic pipelines. It could help laboratories prioritize thousands of detected variants, shorten time to diagnosis, and reduce the number of unresolved cases — which remains one of the biggest challenges in clinical genomics today.

    Related Posts

    News

    White House accuses China of stealing US AI models on an “industrial scale”

    April 24, 2026
    News

    The end of traditional server rooms: Google unveils Virgo Network – a massive network designed exclusively for AI

    April 24, 2026
    News

    75% of Google’s new code is created by AI. The company announces a new era of programming.

    April 23, 2026
    Read more

    Why Haven’t Surgical Robots Taken Over Operating Rooms Yet?

    April 16, 2026

    From AI Picking to Robots by Subscription: How Industrial Robotics Is Changing

    April 15, 2026

    Sex toys got an upgrade. The kitchen didn’t. Maria Kardakova wants to fix that

    April 10, 2026
    Demo
    X (Twitter) Instagram Threads LinkedIn Reddit
    • NEWS
    • ANALYTICS
    • INTERVIEWS
    • ABOUT US
    • EDITORIAL BOARD
    • EVENTS
    • CONTACT US
    • ©2026 2Digital. All rights reserved.
    • Privacy policy.

    Type above and press Enter to search. Press Esc to cancel.