Related Experiment Video
Updated: Jul 2, 2026

Identification and Classification of Position-specific GABAA Receptor Subunit Missense Variants for Their Role In Hippocampal Pyramidal Neurons
Published on: June 6, 2025
From Pathogenicity to Mechanism: A Variant Interpretation Framework for Monogenic Epilepsy.
1Universitat de Barcelona, Barcelona, Spain.
SeizeVar integrates pathogenicity and mechanism to prioritize epilepsy variants of uncertain significance, aiding clinical interpretation. This tool provides a mechanism-annotated list, supporting expert curation and accelerating genetic diagnosis.
Area of Science:
- Genomics and Bioinformatics
- Computational Biology
- Epilepsy Genetics
Background:
- High accuracy of pathogenicity predictors in ClinVar is limited for monogenic epilepsy variants of uncertain significance (VUS).
- Current predictors lack direction-of-effect and scalable mechanism-to-treatment mapping, hindering clinical utility.
- Existing VUS backlogs impede efficient genetic diagnosis and therapeutic development in epilepsy.
Purpose of the Study:
- To develop SeizeVar, a framework integrating pathogenicity and functional mechanism prediction for epilepsy VUS.
- To provide a mechanism-annotated prioritization list to support expert variant curation.
- To establish a community benchmark for reclassification-aware evaluation of VUS.
Main Methods:
- SeizeVar combines a random forest and ESM-2 LoRA cross-attention pathogenicity head with a gain-versus-loss-of-function mechanism classifier.
- A deterministic sodium-channel mechanism-direction rule was incorporated.
- The framework was trained on a 49-gene epilepsy panel and evaluated on multiple held-out cohorts and an external functional cohort.
Main Results:
- SeizeVar's mechanism head achieved an AUROC of 0.736 with 100% panel coverage, comparable to specialized tools.
- General pathogenicity predictors showed significantly lower mechanism prediction performance (AUROC ≤ 0.62).
- Applied to 29,293 epilepsy VUS, SeizeVar identified 4,708 likely pathogenic candidates, with 1,500 sodium-channel variants receiving predicted mechanism-direction labels (679 LoF-leaning, 821 GoF-leaning).
Conclusions:
- SeizeVar effectively integrates pathogenicity and mechanism prediction, offering a novel approach to VUS prioritization in epilepsy.
- The tool provides a mechanism-annotated list, enhancing the efficiency and direction of expert variant curation.
- Prospective clinical validation of SeizeVar predictions is warranted, with the Reclassified-VUS benchmark facilitating future evaluations.
Related Concept Videos
Principles of Pharmacogenetics: Types of Genetic Variants
Encephalitis ll: Pathophysiology
Epilepsy and Seizures: Overview
Various factors can trigger epilepsy, including genetic factors, brain damage, metabolic causes, and unknown etiology. Diagnosis of epilepsy involves electroencephalography (EEG), which...
Epilepsy ll: Types
Epistasis Analysis
Antiepileptic Drugs: Modulators of Neurotransmitter Release Mediated by SV2A Protein
SV2A is a transmembrane glycoprotein located predominantly in the brain, modulating the release of neurotransmitters for neuronal communication. Both levetiracetam and brivaracetam exhibit a high affinity for...

