Rapid identification of genomic structural variations with nanopore sequencing enables blood-based cancer monitoring

Somatic genomic structural variations (SVs) are promising personalized biomarkers for sensitive and specific detection of circulating tumor DNA (ctDNA) in liquid biopsies. However, affordable and fast identification of such SV biomarkers is challenging, which hinders routine use in the clinic. Here, we developed a novel approach - termed SHARC - for rapid discovery of somatic SVs as personalized tumor biomarkers. SHARC combines low coverage cancer genome sketching by using Oxford Nanopore sequencing with random forest classification and a dedicated filtering pipeline to enrich for somatic SVs. Our method leverages the real-time and long-read capabilities of nanopore sequencing to identify somatic SV breakpoints at nucleotide resolution from a tumor tissue biopsy within three days. We applied SHARC to tumor samples of high-grade ovarian and prostate cancer and validated on average ten somatic SVs per sample with PCR mini-amplicons. An accompanying method for SV breakpoint detection from liquid biopsies was devised based on digital PCR, enabling detection of cancer in a quantitative manner. Using this method, we retrospectively monitored treatment response in patients with metastatic prostate cancer. Our work demonstrates that SHARC forms a universal framework for rapid development of personalized biomarker assays for blood-based monitoring of any cancer type.