Mefford, H. | bioRxiv Science

bioRxiv2025

EpiPred: A gene-specific machine learning model for classifying missense variants in the epilepsy-related gene STXBP1

Missense variants in the STXBP1 gene are a frequent cause of early-onset developmental and epileptic encephalopathies and related neurodevelopmental disorders, but the clinical interpretation of these variants remains a major challenge. Most reported STXBP1 missense variants are classified as variants of uncertain significance (VUS), complicating diagnosis, counseling, and patient eligibility for precision therapies. Here, we developed EpiPred, a gene-specific machine learning classifier that predicts the pathogenicity of STXBP1 missense variants by integrating computational features with empirical evidence from cellular assays. Trained on a curated set of pathogenic and benign variants, EpiPred outperformed leading global prediction tools in accuracy, sensitivity, and specificity. We validated the models predictions using functional assays that measure protein abundance, solubility, stability, and interaction with the SNARE complex partner syntaxin 1. These biochemical readouts aligned closely with model outputs and enabled reclassification of several likely misdiagnosed variants. We deployed EpiPred in an interactive web application that allows clinicians, researchers, and patients to explore predictions for all possible missense variants in STXBP1. Our approach illustrates the power of gene-specific predictive modeling combined with experimental validation to improve variant interpretation and diagnostic resolution. By identifying likely pathogenic STXBP1 variants, including those that may respond to emerging therapies such as molecular chaperones, EpiPred supports more precise genetic diagnoses and offers a generalizable framework for other clinically relevant genes in neurological disease. ONE SENTENCE SUMMARYEpiPred improves STXBP1 variant interpretation, enabling precision genetic diagnoses and promoting access to targeted precision therapies

genetics↗

bioRxiv2021

Pathogenic MAST3 variants in the STK domain are associated with epilepsy

ObjectiveThe MAST family of microtubule-associated serine-threonine kinases (STK) have distinct expression patterns in the developing and mature human and mouse brain. To date, only MAST1 has been associated with neurological disease, with de novo variants in individuals with a neurodevelopmental disorder, including a mega corpus callosum. MethodsUsing exome sequencing we identify MAST3 missense variants in individuals with epilepsy. We also assess the effect of these variants on the ability of MAST3 to phosphorylate the target gene product ARPP-16 in HEK293T cells. ResultsWe identify de novo missense variants in the STK domain in 11 individuals, including two recurrent variants p.G510S (n=5) and p.G515S (n=3). All 11 individuals had Developmental and epileptic encephalopathy, with 8 having normal development prior to seizure onset at < 2 years of age. All patients developed multiple seizures types, while 9/11 had seizures triggered by fever and 9/11 had drug-resistant seizures. In vitro analysis of HEK293T cells transfected with MAST3 cDNA carrying a subset of these patient-specific missense variants demonstrated variable but generally lower expression, with concomitant increased phosphorylation of the MAST3 target, ARPP-16, compared to wildtype. These findings suggest the patient-specific variants may confer MAST3 gain-of-function. Moreover, single-nuclei RNA sequencing and immunohistochemistry shows that MAST3 expression is restricted to excitatory neurons in the cortex late in prenatal development and postnatally. InterpretationIn summary, we describe MAST3 as a novel epilepsy-associated gene with a potential gain-of-function pathogenic mechanism that may be primarily restricted to excitatory neurons in the cortex.

genetics↗