Abstract 2704: Resolving clone-and haplotype-specific copy number variation and DNA methylation in heterogeneous tumors with nanopore sequencing

Sergey Aganezov,John Beaulaurier,Eoghan Harrington,Sissel Juul

DOI: https://doi.org/10.1158/1538-7445.am2022-2704

IF: 11.2

2022-06-17

Cancer Research

Abstract:Large-scale somatic genomic copy number variations (CNV) accumulate during cancer progression, resulting in a tumor comprised of collections of cells, or clones, with distinct CNV profiles. Untangling intra-tumor heterogeneity and inferring clone- and haplotype-specific CNV profiles is important for cancer research and can help inform treatment. We present a computational workflow to infer clone- and haplotype-specific cancer CNV profiles by processing long nanopore reads obtained with high-throughput bulk sequencing of a tumor and matching normal sample. The workflow first identifies and phases heterozygous germline single nucleotide polymorphisms (SNPs) in the normal sample. Nanopore reads from the tumor sample are then haplotagged with presence/absence of the phased germline SNPs. Both the overall and the haplotype-specific read counts from the tumor are then tallied over fixed-size bins tiled across the reference genome. Finally, we use the state-of-the art HATCHet matrix factorization algorithm to process the total- and allele-specific read counts and get integer clone- and allele-specific copy number profiles, as well as the clonal cellular fractions of the tumor sample. To test the proposed approach, we have nanopore sequenced to 100x average read-depth coverage a COLO829 tumor and a COLO829BL matching normal cell line. Previous NGS-based bulk and single-cell analyses of the COLO829 cell-line have revealed it to be highly aneuploid and heterogeneous. We show that the proposed nanopore-based workflow identifies clone- and haplotype-specific cancer CNV profiles in concordance with previously published results, with regions of heterogeneity in full agreement with earlier bulk and single-cell studies. Inferred CNV profiles and their clonal fractions are further supported by observed allele-frequencies of somatic SNPs. We demonstrate the stability of the obtained results across lower tumor sample sequencing coverage levels, with CNVs remaining consistent down to 40x tumor coverage, putting the proposed approach on par with the industry standard NGS-based experiments. Notably, because of the unique ability of long nanopore reads to retain single-molecule methylation signals, we were further able to identify haplotype-specific differentially methylated regions both within the tumor sample, as well as in a tumor vs normal comparison, thus shedding light on acquisition/loss of DNA modifications during tumor growth. These results show how nanopore sequencing can be used to resolve some of the complexity that characterizes structurally aberrant heterogeneous cancer samples, while also revealing the previously inaccessible dimension of haplotype-specific tumor methylation. Oxford Nanopore Technologies products are not intended for use for health assessment or to diagnose, treat, mitigate, cure, or prevent any disease or condition. Citation Format: Sergey Aganezov, John Beaulaurier, Eoghan Harrington, Sissel Juul. Resolving clone-and haplotype-specific copy number variation and DNA methylation in heterogeneous tumors with nanopore sequencing [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl) nr 2704.

oncology