Authors :
Presenting Author: Rosita Ramirez Ventura, BS – Mass General Brigham, Harvard Medical School
Sattar Khoshkhoo, MD – Mass General Brigham, Harvard Medical School
Katherine Brown, BS – Mass General Brigham, Harvard Medical School
Michael B Miller, MD, PhD – Mass General Brigham, Harvard Medical School
John Rolston, MD, PhD – Brigham and Women's hospital, Harvard Medical School
Rationale:
Brain-restricted somatic variants have provided critical insights into the genetic underpinnings of focal epilepsy. These variants arise during development and are confined to a subset of cells, resulting in low variant allele fractions (VAFs) that challenge detection. Traditionally, their discovery has relied on brain tissue from resective epilepsy surgery. However, with the increasing use of non-resective approaches such as neuromodulation and Laser Interstitial Thermal Therapy (LITT), access to tissue is limited. Moreover, emerging evidence suggests that somatic genotype influences surgical outcome, underscoring the need for novel methods that detect pathogenic variants before or without resection.
Cells and nuclei adherent to stereo-electroencephalography (sEEG) electrodes—a procedure used to localize seizure foci in patients with medically refractory epilepsy—offer a potential source of patient-specific DNA. However, the predominance of blood- and immune-derived cells complicates detection of low-VAF somatic variants. Additionally, due to the low input DNA obtained from sEEG electrodes, prior methods have relied heavily on extensive pre-amplification before targeted sequencing. This step introduces artifacts that complicate the detection of true low-frequency somatic variants and hinders reliable de novo variant discovery.Methods:
To overcome these limitations, we developed a brain cell enrichment strategy using fluorescence-activated cell sorting (FACS) to deplete blood-derived cells from sEEG electrode-adherent material. To improve variant calling accuracy, we adapted the single-cell duplex sequencing method Multiplexed End-Tagging Amplification of Complementary Strands (META-CS) and made key modifications for our application 1) optimized the protocol to accommodate higher DNA inputs from hundred of cells, 2) improved cell lysis for DNA extraction-free library preparation, and 3) performed targeted capture from META-CS libraries.Results:
We found that only ~30% of sEEG electrode-adherent cells are brain-derived and confirmed enrichment of neuronal and glial populations using single-nucleus RNA sequencing. We validated our workflow by applying targeted META-CS to sorted nuclei from brain samples harboring known somatic variants across a range of VAFs. We then successfully applied the method to sEEG-derived cells, generating high-quality targeted duplex sequencing libraries without any preceding DNA amplification step.Conclusions:
To our knowledge, this work represents the first successful application of targeted duplex sequencing to brain-derived material obtained from sEEG electrodes. Our approach offers a sensitive, scalable, and minimally invasive platform for detecting pathogenic somatic variants in epilepsy prior to resective surgery, with the potential to expand molecular diagnostics and improve care for a broader patient population.Funding:
Blue Sky Award, Epilepsy Foundation of New England