Precision genome profiling for pediatric leukemia with a software-controlled enrichment method using nanopore sequencer.

Abstract

10027 Background: The real-time nature of nanopore sequencing allows for simultaneous basecalling of DNA sequences during sequencing. This unique capability enables adaptive sampling (AS), a software-controlled targeted sequencing method. AS is notable for its ability to rapidly and flexibly enrich multiple genes with long fragment reads and detect various modalities of genetic aberrations in a single run. However, the feasibility and significance of AS for cancer have not been previously reported. Methods: We performed AS on a GridION sequencer using samples from 28 pediatric leukemia patients (10 acute myeloid leukemia, 13 B-cell acute lymphoblastic leukemia, and 5 T-cell acute lymphoblastic leukemia). Target regions were comprised of 466 genes associated with hematologic malignancies and included 30-kilo base pairs of flanking regions. After 3 days of sequencing, single-nucleotide variants (SNVs), structural variations (SVs), and copy number variations (CNVs) were determined. In 21 samples, variant calling accuracy was evaluated using short-read-based whole genome sequencing (WGS). Results: In the on-target regions, mean depth was 21.0× and N50 was 11,191 bps at the median. Of the 13 samples with genetic alterations previously detected through clinical testing for diagnostic classifications, all were reconfirmed by AS. Detecting DUX4 rearrangement needed an additional analysis pipeline because of its structural complexity. Among the 15 remaining samples whose genetic alterations had not been determined by clinical diagnostic testing, AS identified putative driver aberrations in 14 samples. All of these genomic abnormalities were structural variations (SVs) or copy number variations (CNVs), most of which were not encompassed within the genes or regions targeted in clinical diagnostic testing. Detected SVs were described with accurate genomic breakpoints, enabling the efficient detection of gene rearrangements and deletions of the entire region of an exon or a gene. Regarding CNVs, both chromosomal-level CNVs, such as high-hyperdiploid, and focal CNVs, such as CDKN2A deletion, were detectable using genome-wide low-coverage reads in the off-target regions. In the other case with acute myeloid leukemia in which genetic abnormalities were not detected either by clinical testing or AS, WGS identified an NPM1 frameshift deletion located outside of known hotspots. Among variants identified by WGS, AS detected 60.9% of the SNVs, 17.6% of the small indels, and 89.2% of the SVs, with a tendency of poor detection efficiency for variants with low variant allele frequency. Conclusions: AS provides a rapid and precise description of the genomic profile of pediatric leukemia, particularly advantageous for identifying SVs and CNVs, which are difficult to detect by capture-based short-read sequencing.