PO.MCB08.03 · 分子与细胞生物学
针对FFPE肿瘤样本的新流程,通过长读长测序实现结构变异检测和体细胞突变分相的精简解决方案
A new workflow for FFPE tumor samples enables a streamlined solution for structural variant detection and phasing of somatic mutations through long read sequencing
作者与单位 Authors & Affiliations
摘要 Abstract
中文摘要
引言:HiFi长读长测序可对500 bp至25,000 bp长度的片段提供高度准确的读长。这些长读长提供了序列长度的无偏测量,并能实现结构变异检测、分相和直接甲基化分析。然而,FFPE组织对HiFi测序一直较为困难,因为提取常产生短的、损伤的片段,且固定过程使DNA交联,从而破坏长读长测序。
方法:我们应用Covaris truXTRAC FFPE提取技术,通过自适应聚焦声波(AFA)从FFPE蜡块中回收长DNA片段,包括长达5000 bp的分子。为最大化测序效率,我们开发了一种基于Kinnex的文库制备方法,将多个FFPE片段串联成更长的分子。Kinnex涉及PCR,会去除甲基化信息但保留序列质量和连续性。FFPE版Kinnex流程每个样本需要50 ng的输入DNA。
结果:使用该方法对脑、肾和子宫肿瘤FFPE样本进行了测序,每个样本生成超过1亿条HiFi读长,平均读长为750-1,500 bp。所得数据显示出高质量读长和跨基因组一致的变异检测。变异调用每个样本检测到>11,000个结构变异和>510万个小型变异,其中60%的变异被分相为单倍型。
结论:将Covaris长片段提取与Kinnex文库串联相结合,使FFPE样本与HiFi长读长测序兼容。对从Covaris提取的DNA制备的样本进行HiFi测序,从不同组织类型的FFPE样本中生成了高产量测序文库,在不同DNA质量和组织细胞密度下表现出一致的性能。该流程能够对存档的肿瘤基因组进行全面分析,为来自临床相关样本的基因组变异和片段特征提供更完整的视图。
查看英文原文 English abstract
INTRODUCTION: HiFi long-read sequencing provides highly accurate reads of fragments from 500 bp to 25,000 bp in length. These long reads offer an agnostic measure of sequence length and enable structural variant detection, phasing, and direct methylation analysis. However, FFPE tissue has been difficult for HiFi sequencing because extraction often yields short, damaged fragments and fixation crosslinks DNA which disrupts long-read sequencing.
METHODS: We applied the Covaris truXTRAC FFPE extraction by Adaptive Focused Acoustics (AFA) to recover long DNA fragments from FFPE blocks, including molecules up to 5000 bp. To maximize sequencing efficiency, we developed a Kinnex-based library preparation that concatenates multiple FFPE fragments into longer molecules. Kinnex involves PCR, which removes methylation information but preserves sequence quality and contiguity. The Kinnex for FFPE workflow requires 50 ng of input DNA per sample.
RESULTS: Brain, kidney, and uterine tumor FFPE samples were sequenced using this approach, generating more than 100 million HiFi reads per sample with mean read lengths of 750-1,500 bp. The resulting data showed high-quality reads and consistent variant detection across the genome. Variant calling detects >11,000 structural variants and >5.1M small variants per sample, with 60% of variants phased into haplotypes.
CONCLUSIONS: Combining Covaris long-fragment extraction with Kinnex library concatenation makes FFPE samples compatible with HiFi long-read sequencing. Using HiFi sequencing on samples prepared from Covaris-extracted DNA generated high yield sequencing libraries from FFPE samples across diverse tissue types demonstrating consistent performance across varying DNA quality and tissue cellularity. This workflow enables comprehensive analysis of archived tumor genomes, offering a more complete view of genomic variation and fragment characteristics from clinically relevant samples.
利益披露 Disclosure
C. Conner, None..
I. McLaughlin, None..
J. Lake, None..
D. Lee, None..
H. Ferrao, None..
G. Endress, None..
U. Thomann, None..
M. Werner, None..
L. Beker, None.