Using ultra-long-read Oxford Nanopore sequencing to detect complex structural variants in leukaemia | LC 25

Biography

Steven Hair completed his PhD in Aberdeen in molecular biology before moving to Newcastle University to work as a research associate and bioinformatician. He has a particular interest in the genomics of cancer, with a current focus on childhood leukaemia and how long-read sequencing technology can improve both our understanding and diagnosis of these diseases.

Abstract

Many cancers are characterised by genetic abnormalities such as large somatic structural variants (SVs). Many clinically relevant SVs are challenging to detect due to the variable repetitive sequences in which they are located. Oxford Nanopore ultra-long sequencing circumvents the issues of sequencing in repetitive elements thanks to the length of individual reads, therefore allowing for the characterisation of difficult genomic regions and facilitating the detection of SVs in these regions. High molecular-weight DNA was extracted from 15 bone marrow samples from leukaemia patients with known genetic abnormalities. Whole-genome sequencing was performed using the Oxford Nanopore Ultra-Long DNA Sequencing Kit (SQK-ULK114) and a PromethION device. Data was processed through a custom bioinformatics pipeline, with alignment to T2T-CHM13v2.0 genome, SV calling using Severus, and copy number variation calling using Wakhan. The libraries had an average N50 of 76 kb (55—89 kb). SV calling identified an average of 17,244 SVs (16,401–18,107) per sample. The genetic abnormalities IGH::DUX4, iAMP21, IGL::MYC, IGH::MYC, and high hyperdiploidy were all detected using this method. The breakpoints of translocations were resolved to base pair level in all samples and methylation profiling identified hypomethylation of DNA across translocated genes, particularly those within the immunoglobulin (IG) loci. This study demonstrates the potential of ultra-long sequencing as a tool for detecting and characterising complex SVs in cancer. The samples sequenced here represent challenging to characterise SVs, mostly involving the IG loci that can be difficult to detect by standard testing. This study successfully detected and characterised all relevant SVs, along with additional variants that require further investigation.