Evaluation of Long-Read Genome Sequencing for Genomic Profiling of Myeloid Cancers

Abstract

Whole-genome sequencing (WGS) is a comprehensive approach for the genomic evaluation of acute myeloid leukemia (AML) and myelodysplastic syndromes (MDS). We recently described a streamlined tumor-only WGS assay (ChromoSeq) that uses Illumina short-read sequencing with targeted analysis to detect the full range of clinically relevant somatic mutations. Here we sought to determine the performance of this targeted analysis approach using long-read sequencing data from Oxford Nanopore Technologies and Pacific Biosciences. Samples from 26 patients with AML and MDS were sequenced to a mean of 52× coverage. Head-to-head comparison of reportable somatic variants to standard WGS revealed more than 96% recall and 91% precision for single nucleotide variants for both long-read platforms. Performance was lower for insertion/deletions (66% recall and 42% precision), especially in regions with few phased reads that facilitate accurate variant detection. The long-read platforms were 95% accurate for copy number calls, and they detected all recurrent structural variants with no false-positive findings. In addition, long reads properly identified intronic insertions near repetitive elements that were incorrectly identified as interchromosomal structural rearrangements by standard WGS. These results indicate that targeted, tumor-only analysis of long-read sequence data is a feasible approach for the genomic evaluation of myeloid cancers, and they show the utility of incorporating variants discovered via long-read sequencing to improve variant interpretation in short-read WGS.