Focal dystonia gene hunt gets a new entry point: DystoniaGeneNet flags familial MRI signatures, then three new genes

Genetic liability for familial isolated focal dystonia can be read from routine structural brain MRI and used as an entry point for gene screening. Earlier genetic studies were constrained by incomplete penetrance and small family sizes, with fewer than five genes linked to isolated focal dystonia and a substantial share of families still unexplained. DystoniaGeneNet was trained on 40 patients with familial laryngeal dystonia and 40 healthy controls, then correctly classified 18 of 21 independent familial cases, or 85.7%, using a biomarker centred on the genu of the corpus callosum and the bilateral inferior temporal gyrus.

Model output was written directly into variant prioritization and segregation calls, turning an imaging judgement of familial status into testable candidate genes. The usual route selects analysable individuals from clinical phenotype and pedigree structure alone, often leaving non-manifesting carriers out of segregation analysis. Among 214 individuals with whole-exome data, labelling one unaffected relative as a predicted non-manifesting carrier cut six candidates in one family to a single likely pathogenic variant; findings were then searched in another 738 independent exomes, a hospital biobank of 892 cases and 2,475 controls, and All of Us with 1,343 cases and 6,840 controls.

Three new genes place a molecular explanation for dystonia at the extracellular matrix around synapses. RAP1GAP, TNN and ITSN2 had no disease association in OMIM, and isolated focal dystonia lacked an accepted molecular pathway. Blood from the carrier of a RAP1GAP splice variant showed near-complete loss of that transcript, and carriers of an ITSN2 splice variant and a missense variant showed reduced expression across 74 blood RNA profiles; with those three genes excluded, rare nonsynonymous variants in 82 brain-expressed matrisome genes were still associated with dystonia (P = 0.012), with an empirical P = 0.009 against 10,000 random gene sets.

The demonstrated scope is adult-onset isolated focal dystonia cohorts: the model was trained on patients with familial laryngeal dystonia and controls, then applied to familial cases, sporadic patients and unaffected relatives, with genetic findings drawn from whole-exome data, a hospital biobank and European-ancestry participants in All of Us.

The authors note that the discovered variants also appear in non-neurological biobank controls and population databases, where they are rare or ultrarare, so these genes likely show reduced penetrance as well; sample sizes suit family segregation and monogenic discovery, while modifier, common-variant and polygenic-risk, and oligogenic-burden models still need exploration in larger cohorts.