OC48\u2005Re-interrogation of whole exome sequencing data in developmental epileptic encephalopathies

Abstract

Introduction The severe epilepsies of infancy and childhood are a heterogeneous group of severe epilepsies characterised by several seizure types, where the epileptic activity in addition to the seizures contributes to cognitive impairment or regression. They account for a significant proportion of the refractory epilepsies and are usually associated with poor outcome.1 The tern developmental epileptic encephalopathies (DEE) is now the preferred term for this group of children. It may be the result of a specific congenital or acquired structural brain lesions, metabolic disorders, chromosomal abnormalities, copy number variants or single-gene defects. Next-generation sequencing (NGS) includes gene panels, whole-exome sequencing (WES) and whole-genome sequencing. The reported rates of diagnosis in DEE using NGS technology ranges from 10–100%. We previously reported a cohort of 50 patients who underwent single research WES for investigations of DEE. The yield at the time of publication was 22% (11 known epilepsy gene, 1 candidate gene).2 38 patients remained undiagnosed. Since then, the number of new genes reported with DEE continues to expand and the technology improved to aid interpretation of variants. Therefore, we reanalysed WES data, with the addition of parental samples for trio analysis, to enable data interpretation and identification of pathogenic disease-causing variants. Methods Re-analysis of WES data, single (proband only) or trio (proband and parents)WES, if parental samples were available. Results We identified a genetic cause in 25 individuals in the cohort; 22 pathogenic variants in DEE genes, 3 candidate genes, increasing the diagnostic yield to 50%. With re-analysis, we identified 10 pathogenic variants (CDKL5, KCNA2, NRXN1, PRODH, RELN, RHOBTB2, SCN1A, SLC1A4, SMC1A), 1 candidate gene (NAPB) and 1 variant of uncertain significance (GRIN2A). A number of genes had not been identified at the time of initial analysis, including RHOBTB2, SLC1A4, SMC1A. Two mosaic variants in CDKL5 and SCN1A were identified with trio WES analysis and reducing read depth filter to 15, previously set at 20. Discussion This study highlights the importance of the re-interrogation of WES data for newly discovered genes. Trio WES had a higher diagnostic yield (50% compared to 22%) in keeping with previous studies. Trio WES is effective for the identification of de novo variants and aids in the interpretation of variants. Reducing read depth filter can aid the identification of mosaic variants, increasing reported to be important in DEE.