Variant recurrence in neurodevelopmental disorders: the use of publicly available genomic data identifies clinically...

Next-generation sequencing has revealed the major impact of de novo variants (DNVs) in developmental disorders (DD) such as intellectual disability, autism, and epilepsy. However, a substantial fraction of these predicted pathogenic DNVs remains challenging to distinguish from background DNVs, notably the missense variants acting via nonhaploinsufficient mechanisms on specific amino acid residues. We hypothesized that the detection of the same missense variation in at least two unrelated individuals presenting with a similar phenotype could be a powerful approach to reveal novel pathogenic variants.

We looked for variations independently present in both our database of >1200 solo exomes and in denovo-db, a large, publicly available collection of de novo variants identified in patients with DD.

This approach identified 30 variants with strong evidence of pathogenicity, including variants already classified as pathogenic or probably pathogenic by our team, and also several new variants of interest in known OMIM genes or in novel genes. We identified FEM1B and GNAI2 as good candidate genes for syndromic intellectual disability and confirmed the implication of ACTL6B in a neurodevelopmental disorder.

Annotation of local variants with denovo-db can highlight missense variants with high potential for pathogenicity, both facilitating the time-consuming reanalysis process and allowing novel DD gene discoveries.