AI Tool Offers New Hope for Pinpointing Harmful Epilepsy Genetic Variants

For families navigating the complexities of rare childhood epilepsies, the journey often involves significant uncertainty, particularly when genetic testing reveals “variants of uncertain significance” (VUS). These ambiguous findings can leave clinicians and families in limbo, unable to definitively determine if a genetic difference is the cause of a child's condition or merely a benign variation. This diagnostic challenge can have profound implications, including precluding patients from participating in emerging gene-targeted therapies.

A significant development from Northwestern Medicine scientists and collaborating institutions offers new hope. They have developed an artificial intelligence tool, EpiPred, designed to more accurately identify whether specific genetic variants associated with a severe form of childhood epilepsy, developmental and epileptic encephalopathies, are indeed disease-causing. This innovation, detailed in a study published in the Journal of Clinical Investigation, represents a crucial step toward connecting patients with potentially life-changing precision treatments.

Bridging the Diagnostic Gap with AI

The research, led by Gemma Carvill, PhD, adjunct associate professor at Northwestern University Feinberg School of Medicine, focuses on the STXBP1 gene. This gene is commonly linked to developmental and epileptic encephalopathies, a group of neurodevelopmental disorders characterized by early-onset seizures and significant developmental delays. While genetic testing can identify variants in STXBP1, many are currently classified as VUS, meaning there isn't enough information to determine their clinical significance.

As Dr. Carvill explains, this uncertainty has direct consequences for affected families. “If you send off for clinical genetic testing, you get a variant of uncertain significance in this gene, it essentially precludes you from any future gene targeting therapies,” she states. A non-definitive diagnosis means patients might not be eligible for new precision therapies specifically designed to target the gene of interest.

To overcome this hurdle, investigators developed EpiPred, a gene-specific machine learning model. This model was trained to predict whether STXBP1 missense variants are likely to be disease-causing. The team then rigorously tested EpiPred's predictions using multiple laboratory assays that measured various protein characteristics, including abundance, stability, solubility, and interaction with other cellular components.

Enhanced Accuracy and Clinical Implications

According to Northwestern Medicine, EpiPred demonstrated superior performance compared to existing global variant prediction tools in terms of accuracy, sensitivity, and specificity. The model's predictions also closely aligned with experimental laboratory findings, enabling investigators to identify several variants that may have been misclassified in previous studies.

“We found some misdiagnoses. In other words, individuals who have been diagnosed with STXBP1-related disorder, but we actually think that that is a false diagnosis,” Dr. Carvill notes. Correcting these diagnoses is becoming increasingly vital as gene-targeted therapies progress toward clinical trials. “Being accurate in these diagnoses is really important as we move towards these precision therapies,” she emphasizes.

The research team further examined hundreds of known variants and predicted the impact of every possible STXBP1 missense variant. Their findings suggest that approximately one-quarter of variants currently classified as uncertain may actually be disease-causing. This discovery could potentially lead to definitive diagnoses for more patients, making them eligible for future treatments.

One of the study's most significant findings, according to Dr. Carvill, is the model's ability to distinguish harmful variants with very high accuracy. This capability could empower clinicians with clearer information, streamlining diagnostic processes and informing treatment pathways.

Accessibility and Future Directions

In a move to maximize the impact of their work, the investigators have made EpiPred publicly available through an interactive web application. This platform allows clinicians, scientists, and families to explore predictions for STXBP1 variants, fostering a more informed and collaborative approach to diagnosis and care.

The research, conducted as part of a National Institutes of Health-funded collaboration involving Northwestern Medicine, the University of Michigan, Weill Cornell Medicine, and St. Jude Children’s Research Hospital, combined computational approaches with cellular and animal model systems. This multidisciplinary collaboration was key to improving the interpretation of epilepsy-related genetic variation.

Looking ahead, Dr. Carvill and her team are developing higher-throughput approaches to evaluate thousands of variants simultaneously. Their goal is to dramatically expand the scale of their initial work, which evaluated roughly two dozen variants using four separate experimental assays. “Now we have developed two assays where we can look at every single possible variant that could ever be made in STXBP1,” Dr. Carvill states, explaining that they can assess the impact of over 3,000 variants in this single gene on protein function.

Ultimately, the investigators hope their innovative approach will extend beyond epilepsy to other rare genetic disorders. This framework, combining gene-specific AI models with laboratory validation, holds the potential to improve diagnosis, guide clinical trial enrollment, and advance precision medicine across a broader spectrum of conditions.

Learning and Understanding Complex Medical Advancements

Understanding the intricacies of genetic research and AI applications can be challenging, even for adults. For students curious about how AI is revolutionizing medicine or the fundamentals of genetics, COSMIQ offers a free voice-driven AI tutoring platform. K-12 students can explore complex topics, ask questions, and deepen their understanding of scientific breakthroughs like EpiPred, making advanced concepts more accessible. Similarly, for those preparing for exams that touch on biology or technology, resources like the exam prep directory can provide valuable support.

This pioneering work by Northwestern Medicine and its collaborators highlights the power of interdisciplinary research and advanced computational tools in addressing some of the most pressing challenges in rare disease diagnosis. By providing clearer answers about genetic variants, EpiPred offers a beacon of hope for families affected by severe childhood epilepsies, paving the way for more precise and effective treatments.