Abstract 6631: High-resolution characterization of FFPE-derived linked-read libraries for cancer genomics

Abstract Deciphering genomic architecture is crucial for identifying novel disease drivers and understanding cancer initiation and progression mechanisms. While short-read sequencing remains the primary tool in cancer genomics, it may miss important structural variants (SVs), especially large chromosomal alterations. A further challenge to cancer researchers is working with FFPE (formalin-fixed, paraffin-embedded) samples, the most common source of archived material, as this sample type is often at risk of nucleic acid fragmentation, DNA crosslinks, and deamination, further complicating variant discovery. We present a Dovetail-FFPE assay, an MNase-HiC-based method compatible with FFPE samples that generates linked-reads to enable improved sensitivity for SV detection. Here, we evaluate the performance of Dovetail-FFPE against a shotgun sequencing library in FFPE and FF (fresh frozen) using a clinical serous ovarian adenocarcinoma (HGSOC) sample. We generated WGS and Dovetail-FFPE libraries at 30x target coverage and utilized BWA for alignment, with Hi-C breakfinder (Dovetail-FFPE) and Manta (WGS) for SV calling. Both Dovetail-FFPE and WGS datasets were analyzed using DeepSomatic for SNV and intel detection, while CNVs were identified using Purple. Dovetail-FFPE analysis revealed four translocations (tlocs) with strong supporting evidence (1;12, 3;12, 3;4, 2;15). These tlocs, also identified in a MicroC-FF sample and validated by PCR and Sanger sequencing, corresponded with copy number changes and occurred at CNV boundaries. In contrast, WGS detected only two of these four tlocs with lower precision. Somatic mutations in coding regions showed high concordance between methods. The Dovetail-FFPE library successfully recalled all pathogenic variants from the truth set (defined as pathogenic/likely pathogenic significance by ClinVar in the FF WGS sample). Copy number analysis demonstrated strong concordance between Dovetail-FFPE and FF-WGS (89%), recalling a heterozygous deletion of chromosome 1p and gains of 1q and 12. Additionally, shotgun, Dovetail-FFPE, and MicroC-FF showed agreement in predicted purity, ploidy, and percentage of genome altered. Our findings demonstrate that Dovetail-FFPE libraries can effectively detect SVs identified in FF libraries and outperform shotgun libraries. Combined with its capability to identify SNVs, indels, and CNVs, linked-read sequencing proves to be a valuable tool for comprehensive genetic variation discovery. Citation Format: Alexander Fortuna, Jon Torchia, Mital Bhatka, Natalie Fredriksson, Daniel Hwang, Ek Danwanichakul, Cory Padilla, J. Zachary Sandborn, Lisa Munding. High-resolution characterization of FFPE-derived linked-read libraries for cancer genomics [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 6631.