Targeted nanopore sequencing with Cas9-guided adapter ligation

Abstract

Despite recent improvements in sequencing methods, there remains a need for assays that provide high sequencing depth and comprehensive variant detection. Current methods^1-4 are limited by the loss of native modifications, short read length, high input requirements, low yield or long protocols. In the present study, we describe nanopore Cas9-targeted sequencing (nCATS), an enrichment strategy that uses targeted cleavage of chromosomal DNA with Cas9 to ligate adapters for nanopore sequencing. We show that nCATS can simultaneously assess haplotype-resolved single-nucleotide variants, structural variations and CpG methylation. We apply nCATS to four cell lines, to a cell-line-derived xenograft, and to normal and paired tumor/normal primary human breast tissue. Median sequencing coverage was 675× using a MinION flow cell and 34× using the smaller Flongle flow cell. The nCATS sequencing requires only ~3 μg of genomic DNA and can target a large number of loci in a single reaction. The method will facilitate the use of long-read sequencing in research and in the clinic.

Figure 1 -. Method schematic and coverage data

(A) Schematic of Cas9 enrichment operation. ROI = region of interest. First, DNA ends are dephosphorylated, then new cuts introduced with Cas9/guideRNA complex, and nanopore sequencing adaptors are ligated to cuts around the ROI prior to loading the sample on the nanopore sequencer. (B) Coverage plots at the KRT19 gene (enriched area 18kb) in four separate enrichment experiments: GM12878 with a single gRNA on each side (minION); GM12878 with three gRNA on each side (minION); GM12878 with three gRNA on each side (flongle); and fresh tumor tissue with three guideRNAs on each side (minION). (C) Table showing total aligned read count, on-target reads (within 20kb of a guideRNA site), on-target percentage, and median coverage at each of the ten enriched regions.

Figure 2 -. Single Nucleotide Variants

(A) Plot of sensitivity versus coverage using four tools to call single nucleotide variants from enrichment data in GM12878 for a 140kb region containing 174 annotated SNVs (B) Visual representation of high-confidence variants detected by nanopolish in the MinION data from GM12878 for the captured region around TP53, reads phased into homologous alleles using WhatsHap. (C) High-confidence variants identified in primary tissue from a tumor/normal pair, red arrows used to demarcate tumor-specific variants.

Figure 3 -. Methylation Analysis

(A) Read-level plots showing methylation patterns in GM12878 from minION and flongle data at the KRT19 locus. (B) Methylation calls (points) and line plots at the same locus as in (A) showing smoothed (loess) methylation calls from whole genome bisulfite sequencing on the Illumina platform, compared with methylation calls from minION and flongle targeted nanopore sequencing. (C) Haplotype phased methylation calls in primary patient tissue and paired tumor at the KRT19 locus.

Figure 4 -. Structural Variation

(A) Reads around an ~8kb deletion in chromosome 7 present in MCF-7 and MDA-MB-231, and absent in MCF-10A. (B) Coverage on each parental allele in the region of a large (155kb) heterozygous deletion in GM12878. (C) Top: Coverage at the BRCA1 locus from DNA extracted using Circulomics CBB kit. Middle: LINE and SINE components identified by RepeatMasker on each of the BRCA1 allele assemblies. Bottom: Three indels discovered between BRCA1 assemblies not annotated in platinum genome data set for GM128789.