VUS. The story begins with a patient in clinic. A young child with severe epilepsy, carrying a variant in SCN1A, the classic gene for Dravet Syndrome. But the variant is labeled a variant of uncertain significance (VUS). Dravet Syndrome is a clinical diagnosis, and the treatments we have today do not hinge on whether the variant is clearly pathogenic or not. But then we wonder whether a novel precision therapy could be an option, and we look up inclusion criteria and hesitate. Trial frameworks often require a variant to be pathogenic or likely pathogenic, and future precision medicine approaches in routine clinical care may require the same. For this patient, a VUS is a door that does not open. Here lies the quiet revolution in epilepsy genetics that is unfolding in the background: the refinement of variant interpretation itself.
Figure 1. Visualizing the new ACMG/AMP points-based classification system. The diagram shows how different combinations of evidence criteria contribute to variant interpretation under the updated framework. Each bar represents a typical evidence scenario, broken down by individual ACMG/AMP rules such as PVS1 (very strong), PS2–PS4 (strong), PM1–PM6 (moderate), and PP1–PP5 (supporting). Conflicting evidence is indicated by hatched bars, highlighting situations where pathogenic and benign signals pull in opposite directions. Colored background zones indicate the thresholds that define final classifications: Pathogenic (≥10 points), Likely Pathogenic (6–9 points), VUS (0–5 points), Likely Benign (−1 to −6 points), and Benign (≤−7 points, or BA1 as a stand-alone benign criterion). This new quantitative framework turns what once felt like a “word salad” of rules into a transparent and teachable scoring system, marking a quiet revolution in how we approach clinical variant interpretation.
Variant curation as a revolution. In the field of epilepsy genetics, we talk about sequencing technologies, big data, and gene discovery as the milestones that shaped modern epilepsy genetics. But what if the most profound change today is not about finding new genes at all? What if the revolution is happening in the rules that tell us what counts as “pathogenic”? Variant curation criteria, refined gene by gene, are now starting to shift how we interpret the genomic information we already have. This is the revolution you probably have not heard of — but it may soon shape clinical decisions more directly than new discoveries.
The sodium channel test case. Take SCN1A, SCN2A, and their paralogs. For years, we applied the generic ACMG/AMP rules across all genes. The problem was obvious: epilepsy genes do not behave like cancer genes, and sodium channels do not behave like collagen genes. A missense variant in a critical transmembrane segment of a sodium channel carries more weight than one in a flexible tail region, but the standard framework could not capture that nuance. ClinGen expert groups have now refined these criteria—introducing gene-specific cut-offs, phenotype-weighted de novo points, and quantitative thresholds for functional assays. These may sound like technical details, but they change the classification for real patients, shifting variants out of the uncertain zone.
The scoring system explained. At the heart of this shift is a points-based scoring system. Each piece of evidence about a variant—whether it appears de novo, whether it affects a critical domain, whether functional studies show a change—earns a certain number of points. Add enough points together, and a variant moves from uncertain to likely pathogenic, or from likely benign to clearly benign. It is a Bayesian framework hiding under the hood of variant interpretation. The numbers matter: a de novo event with the right clinical phenotype may be worth one point, while a functional assay showing a major channel defect might be worth two. For the first time, there is a quantitative backbone to what used to feel like expert judgment alone.
The alphabet soup of ACMG. But here is the challenge—the language of the ACMG framework can feel like word salad to anyone first encountering it. Terms like PVS1, PM2, PS3, or BP4 are shorthand for specific evidence types, but to learners they can sound like codes without a key. For example, PVS1 refers to a null variant expected to cause loss of function, PM2 is about rarity in population databases, and PS3 points to functional studies showing a damaging effect. Even more confoundingly, criteria can be modified to different strengths, meaning PM6 (assumed de novo, but without confirmation of paternity and maternity), despite including an ‘M’ for moderate, can be applied at any evidentiary level from supporting to very strong. To experienced curators these criteria are second nature, but for students and clinicians they can seem opaque. The new gene-specific refinements help make sense of this alphabet soup—they anchor each criterion in concrete numbers and gene-specific rules. This is what transforms a theoretical framework into a practical tool.
Why it matters now. This revolution is not hypothetical. In March 2024, we changed the ACMG specifications for epilepsy-related sodium channels, which now allow us to interpret variants with a level of precision that directly impacts families. A patch clamp experiment that once gave us “supporting evidence” now has calibrated numeric thresholds. De novo variants in patients with developmental and epileptic encephalopathies (DEEs) contribute points in a weighted system. Mutational hot spots are mapped, not guessed. For the family in clinic, this means that a variant once classified as “uncertain” may finally cross the threshold into “pathogenic” or “likely pathogenic.” And that opens doors—to therapies, to trials, to answers.
What you need to know. The biggest revolution in epilepsy genetics is not a new technology or a new gene. It is the quiet refinement of the rules we use to interpret the data already in hand. Variant curation criteria are becoming gene- and disease-specific, moving us past the era of broad-brush ACMG rules. This shift means that a patient with an SCN1A variant of uncertain significance today may soon have a clear answer tomorrow. And in the landscape of precision therapies, that difference is everything.
