94P - Clinical utility of long read sequencing for comprehensive analysis of cancer patient genomes

Background

Long read sequencing (LRS) technologies offer a number of advantages for the clinical characterisation of cancer genomes over the short read sequencing (SRS) technologies currently employed in clinical practice. Genomics England in conjunction with NHSE have initiated a pilot programme to establish the clinical utility of Oxford Nanopore® (ONT) LRS for a subset of cancer patients.

Methods

We have developed an approach to generate high depth, paired (50X tumour, 25X normal) whole genome sequences with LRS and a bioinformatics pipeline to call somatic variants. Our pipeline calls Single Nucleotide Variants (SNVs) and small insertions and deletions (indels) with ClairS, Structural Variants (SVs) with Severus and Copy Number Variants (CNVs) with Purple. We have generated sequencing data for over 100 patients through the NHS Genomic Lab Hubs, and approximately 400 patients as part of our research cohorts, focussing on leukaemia, sarcoma and brain tumour patients.

Results

Our ONT cancer analysis pipeline can detect clinically relevant SNVs, CNVs and SVs with precision and recall exceeding 90%. LRS offers a superior view of complex structural variation, for example phasing of SV breakpoints can distinguish multiple events occurring in cis or trans phase. Furthermore, variants in hard-to-map regions of the genome, such as repetitive sequences, that are inaccessible to SRS technology can be analysed using LRS, allowing detection of additional clinically relevant variants. Characterisation of tumour-specific DNA methylation patterns allows for classification of brain tumour subtypes with machine learning tools, with similar accuracy to data generated from methylation arrays. Our pipeline can also detect and characterise methylation of clinically relevant regions, such as the MGMT and BRCA1 promoter regions.

Conclusions

We have demonstrated that cancer genomes can be comprehensively and accurately characterised with ONT sequencing. This data will be made available as a resource for future research. Comparison with the current standard of care testing will be assessed in the context of the clinical impact for each patient group. This will inform recommendations for the use of ONT for standard of care testing.

Editorial acknowledgement

Clinical trial identification

Legal entity responsible for the study

Genomics England.

Funding

Genomics England.

Disclosure

All authors have declared no conflicts of interest.