Deep whole genome sequencing of epileptogenic brain lesions.

Objective

Malformations of cortical development and low-grade epilepsy-associated tumors often cause drug-resistant lesional focal epilepsy (LFE) amenable to surgical treatment. Genetic testing of resected brain tissue can provide diagnostic and mechanistic information and may eventually support personalized treatment approaches. However, systematic >300× whole genome sequencing (GS) of bulk epileptogenic lesions, particularly after nondiagnostic deep exome sequencing and array-based genotyping, has not been evaluated.

Methods

Here, we performed deep GS (>300× coverage) on 33 epileptogenic brain lesions from individuals with LFE, selected for lesions with a high prior probability of a genetic finding, all of whom had negative results for candidate germline or somatic variants in previous deep exome sequencing (ES; >350× coverage) and array-based genotyping.

Results

We identified candidate somatic variants in known LFE-associated genes in seven of 33 participants (21%). Two participants had two candidate variants each. In total, three of 33 participants (9%) had variants that could only be detected by GS: two deletions in TSC2, two deletions in NPRL3, and an exonic inversion in FGFR1. Three individuals had germline polymorphisms that may represent modulators or risk factors. Exploratory mutational signature analysis revealed clocklike signatures and a correlation between mutational signatures, age at seizure onset, and histopathology, indicating an age-related mechanism of mutation.

Significance

Deep GS establishes the feasibility of genome-wide profiling at a depth relevant to low-variant allelic frequency mosaicism and reveals candidate variants missed by ES, particularly complex structural variants. As sequencing costs decrease, GS may complement other technologies in somatic variant detection, enabling more widespread clinical implementation of tissue sequencing in epilepsy surgery. Further studies are needed to address the remaining conceptual and technological challenges in somatic variant detection.