Abstracts

Rationale: Epilepsy is the third most common neurological disorder, affecting 65 million people worldwide. To better understand and treat epilepsy, it is important to identify the genetic causes of the disorder. Approximately two thirds of epilepsies are estimated to have a genetic basis, but in most cases, there is no molecular diagnosis. Some epilepsies can be caused by somatic mosaicism, which occurs as the result of a postzygotic DNA mutation. In the case of somatic mosaicism-associated epilepsy, only a particular region of the brain may carry a pathogenic variant. Such somatic variants have been identified in both lesional and non-lesional epilepsies. To date, the majority of pathogenic somatic mosaic variants identified in individuals with epilepsy have been single nucleotide variants (SNVs) and small insertion-deletions (indels). The role of mosaic structural variants (SVs) such as copy number variants (CNVs) in epilepsy is less well-established. In this study, we aimed to determine whether brain-specific somatic mosaic SNVs/indels and SVs can explain some instances of epilepsy with previously unknown genetic etiology.

Methods: We enrolled individuals who undergo epilepsy surgery, either steroelectroencephalography (SEEG) or resection, and obtain brain tissue and whole blood. We extracted genomic DNA from brain tissue and leukocytes, prepared next generation sequencing libraries using Agilent’s SureSelect XT HS technology, and sequenced to a depth of 500X (brain) and 100X (leukocytes). We used GATK-HaplotypeCaller (GATK-HC) to identify high variant allele frequency (VAF 0.2-0.5) SNVs/indels and compared matched blood-brain pairs to identify brain-specific variants. Low VAF (0.02-0.2) mosaic SNVs/indels were identified using MosaicForecast. SVs are identified from exome sequencing using MrMosaic. Somatic SNVs/indels and SVs will be validated and precise levels of mosaicism identified using custom targeted enrichment probes and ultra-deep sequencing (2000X).

Results: Using GATK-HC and MosaicForecast, we identified ~150-600 high confidence (read depth >30) brain-specific SNVs/indels per individual from deep exome sequencing of 19 matched blood-brain samples. In one individual, we identified a GRIN2B missense variant present at a VAF of 0.24 in resected brain tissue only. CNV calling is ongoing.

Conclusions: This study establishes protocols to identify two major classes of genetic somatic variation in brain tissue of individuals with epilepsy and provides the basis for ongoing studies to determine whether cell-free DNA released from cells in the brain and found in plasma can be used to identify brain-specific variants.

Funding: Please list any funding that was received in support of this abstract.: 1DP2NS111506-01 NIH Director’s New Innovator Award.