PO.MCB08.03 · 分子与细胞生物学

采用纳米孔自适应采样进行单一检测、仅肿瘤样本的体细胞SNV、SV与CNV分析

Single assay, tumor-only, somatic SNVs, SVs, and CNVs profiling using nanopore adaptive sampling

海报缩略图:采用纳米孔自适应采样进行单一检测、仅肿瘤样本的体细胞SNV、SV与CNV分析
编号 3239 展板 4 时间 4/20 02:00–05:00 区域 Section 22 主讲 Sergey Aganezov, PhD
分会场 Genomic Profiling to Understand Cancer Biology
查看 PDF 下载 PDF 🔒 查看 / 下载完整 PDF 需登录并开通下载套餐 · 查看套餐 / 开通 AACR 官方页面

作者与单位 Authors & Affiliations

Sergey Aganezov1, Philipp Rescheneder2, Rory Sinnott3, Sissel Juul1

1Oxford Nanopore Technologies, Inc, New York, NY,2Oxford Nanopore Technologies GmbH, Munich, Germany,3Oxford Nanopore Technologies plc, Oxford, United Kingdom

摘要 Abstract

中文摘要
背景。 仅肿瘤(T-only)检测往往需要在多方面进行权衡:当采用基于扩增子或杂交捕获的富集方法时,可在低等位基因分数(AF)下获得高灵敏度;而当采用全基因组测序时,则可获得全面的结构与拷贝数分析以及整体上更好的易用性。纳米孔自适应采样(AS)通过在测序过程中剔除与用户定义靶标不重叠的读段来富集靶标,从而在靶标区域产生高深度(100-200X)长读段(8-15kb)数据,在非靶标区域产生浅测序(5-15X,500bp)数据。因此,AS无需PCR或基于pulldown的富集,从而简化了实验设置,同时能够在单一检测中检测单核苷酸变异(SNV)、结构变异(SV)和拷贝数变异(CNV)。 方法。 在此,我们评估了仅肿瘤AS在一个高度畸变基因组(COLO829)以及一个由COLO829及其配对正常样本在计算机中构建的、畸变程度较低的"合成"基因组(仅保留少量大规模CNV事件)上的变异检出准确性。性能评估跨越了模拟常见条件的覆盖度/纯度梯度。测序针对一个通用的多区域、覆盖700余个癌症相关基因的panel,并经过扩充以覆盖COLO829中存在的一系列SV类型。SNV和SV通过长读段仅肿瘤流程进行推断。SV包括易位、缺失、重复和插入。全基因组CNV图谱则通过利用非靶标读段从同一批AS数据集中获得。 结果。 SNV。 高(100-200X)的靶标区域深度使得能够在相关AF水平上可靠地回收体细胞SNV,在低至0.05 AF时具有高召回率,在低至0.1-0.2 AF时具有高精确度。这些结果在覆盖度/纯度梯度上保持一致。 SV。 长读段靶标区域数据支持对染色体间和染色体内重排的稳健检测。值得注意的是,即使事件仅部分与靶标区域重叠,也能被回收。 CNV。 大量短的非靶标读段从同一批运行中生成了稳定的全基因组CNV分段。所推断的CNV图谱在覆盖度/纯度梯度上,在高度畸变基因组以及畸变程度较低的合成构建体上均展现出预期的大范围增益与缺失。 结论。 单次仅肿瘤纳米孔AS运行可以(1)通过高靶标覆盖度实现灵敏的SNV检测,(2)提供稳健的SV发现,包括单断裂端在靶标区域内的情形,以及(3)从非靶标信号中获得全基因组CNV图谱。通过简单的软件定义靶向来避免繁琐的panel特异性杂交捕获检测,AS为单一检测的仅肿瘤样本表征提供了一种实用、简化且高度灵活的方法,覆盖了癌症中发现的全部基因组变异范围。
查看英文原文 English abstract
Background. Tumor‑only (T‑only) testing often trades high sensitivity at low allele fraction (AF) - when using amplicon or hybrid-capture based enrichment - against comprehensive structural and copy‑number profiling and overall ease of use, when using whole genome sequencing. Nanopore adaptive sampling (AS) enriches targets by rejecting reads during sequencing if they don't overlap with user‑defined targets, yielding high-depth (100-200X) long read (8-15kb) data on target and shallow (5-15X and 500bp) sequencing off-target. As a result, AS simplifies experimental setup by not requiring PCR or pulldown-based enrichment, while at the same time enabling the detection of single nucleotide variants (SNVs), structural variants (SVs), and copy number variations (CNVs) in a single assay. Methods. Here, we evaluated variant calling accuracy for T‑only AS on a highly aberrant genome (COLO829) and on a “synthetic,” less‑aberrant genome constructed in silico from COLO829 and its matched normal, retaining only a small number of large‑scale CNV events. Performance was assessed across coverage/purity ladders that emulate common conditions. Sequencing targeted a generic multi‑region, 700+ cancer‑gene‑aware panel augmented to cover a range of SV types present in COLO829. SNVs and SVs were inferred with long‑read tumor‑only pipelines. SVs included translocations, deletions, duplications, and insertions. Genome‑wide CNV profiles were derived from the same AS datasets by leveraging off‑target reads. Results. SNVs. High (100-200X) on‑target depth enabled confident recovery of somatic SNVs at relevant AFs, with high recall down to 0.05 AF and high precision down to 0.1-0.2 AF. These results were consistent across coverage/purity ladders. SVs. Long‑read on-target data supported robust detection of inter‑ and intra‑chromosomal rearrangements. Notably, events were recovered even when only partially overlapping with the target regions. CNVs. The high number of short off‑target reads generated stable, genome‑wide CNV segmentations from the same runs. Inferred CNV profiles demonstrated expected broad gains and losses on the highly aberrant genome as well as on the synthetic, less‑aberrant construct across coverage/purity ladders. Conclusions. A single T‑only nanopore AS run can (1) deliver sensitive SNV detection via high on‑target coverage, (2) provide robust SV discovery, including one‑breakend-on-target cases, and (3) yield genome‑wide CNV profiles from off‑target signal. By avoiding laborious panel‑specific hybrid capture assays through simple software‑defined targeting, AS offers a practical, streamlined, and highly flexible approach to single-assay tumor‑only characterization that covers the full breadth of genomic variation found in cancer.
利益披露 Disclosure
S. Aganezov, Oxford Nanopore Technologies, Inc Employment. Oxford Nanopore Technologies plc Stock, Stock Option. P. Rescheneder, Oxford Nanopore Technologies GmbH Employment. Oxford Nanopore Technologies plc Stock, Stock Option. R. Sinnott, Oxford Nanopore Technologies plc Employment, Stock, Stock Option. S. Juul, Oxford Nanopore Technologies, Inc Employment. Oxford Nanopore Technologies plc Stock, Stock Option.

← 返回 AACR 2026 检索