Abstract 6610: Detection of homologous recombination deficiency in targeted long read sequenced tumors using mutational signatures

Abstract Background: Homologous recombination deficiency (HRD) affects cancer types including breast, ovarian, pancreas and prostate and patients with HR deficient tumors can respond to poly ADP-ribose polymerase (PARP) inhibitors and platinum-based therapies. HRD is associated to the loss of either germline or somatic, BRCA1 or BRCA2 and other associated variants. However, a large proportion of tumors do not have an identifiable variant in BRCA1 or BRCA2 and other related genes. Patterns of mutations—Mutational Signatures—of HRD can identify changes in DNA sequences due to defective DNA repair and genomic instability in HRD tumors. Current computational tools to detect HRD are based on whole genome sequencing (WGS) and an assessment of robustness of these approaches for clinical implementation are lacking. Methods: We developed mSigDetect to identify HRD associated mutational signature (SBS3) in breast cancers, sequenced on Oxford Nanopore Technologies (ONT) targeted adaptive sampling technology. We built our mutational signatures classifier based on mutational profiles of 567 breast cancers from the International Cancer Genome Consortium (ICGC) and tested mSigDetect and the targeted panel on synthetic tumors (n=200) and two independent cohorts of breast cancer patients sequenced on Illumina short reads (n=224) and on ONT long reads (n = 38).Using the simulated synthetic tumours, we determined the sensitivity and specificity of our tool. We tested the performance of mSigDetect under different scenarios of tumour cellularity by down sampling and mixing HR deficient cell lines with paired normal lines in a range of 30-90% tumour cellularities. Last, we developed a computational workflow for ONT sequencing, to carry out real-time variant calling and mutational signature identification to reduce the turnaround time of the diagnostic test. Results: With a target region of 5% of the human genome sequenced, mSigDetect could identify HR deficient tumors in both synthetic (sensitivity = 0.73, specificity = 1) and breast cancer patient tumors (sensitivity = 0.81, specificity = 0.99) compared to WGS. We designed mSigDetect to detect SBS3 in a tumor-only setting, without the need of a paired normal sample. Signature detection of SBS3 using mSigDetect greatly outperformed sequencing whole exome targeted regions. Based on cell line mixing simulations, we recommend sequencing at 45x for 30% tumor purity or higher, or 30X for 50% purity. ONT adaptive sampling identifies HR deficient tumors robustly in a clinical setting, based on targeted sequencing of a tumor-only sample. Citation Format: Alvin Wei Tian Ng, Sheryl Li Lynn Wong, Siao Ting Chong, Jieru Ye, Joanne Ngeow. Detection of homologous recombination deficiency in targeted long read sequenced tumors using mutational signatures [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2025; Part 1 (Regular Abstracts); 2025 Apr 25-30; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2025;85(8_Suppl_1):Abstract nr 6610.