Beyond the BAM – revealing new dimensions in clinical research data | LC26

Abstract
Genomic characterisation of tumour cells is an essential component of modern risk-stratified therapy in B-cell precursor acute lymphoblastic leukemia (BCP-ALL). Short-read next-generation sequencing (NGS)–based transcriptome analysis has identified over 25 genomic subtypes. However, implementation of these approaches in low- and middle-income countries (LMICs) is limited by cost, infrastructure, and longer turnaround times. Oxford Nanopore sequencing offers a potential solution by enabling rapid and cost-effective genomic characterisation. Transcriptome sequencing was performed on 94 samples, previously characterised by cytogenetics and immunophenotyping, using orthogonal sequencing platforms: Illumina NextSeq 550 and Oxford Nanopore Technologies (ONT) MinION/PromethION. Illumina gene-expression data was analysed using the molecular diagnosis of acute lymphoblastic leukemia (MD-ALL) algorithm, while ONT data was analysed for lineage and subtype classification using a novel machine-learning algorithm. Comparative analysis demonstrated concordance with conventional cytogenetic and molecular findings in 83% (78/94) of cases using Illumina whole-transcriptome sequencing (WTS) and 84% (79/94) using nanopore WTS, alongside identifying expression-based subtypes like DUX4, PAX5alt, Ph-like, ETV6-RUNX1-like subtypes. The classifier achieved 99% accuracy for lineage prediction and 86% accuracy for genomic subtype prediction. Subtype misclassification occurred in 10/34 samples sequenced on MinION and 4/59 samples sequenced on PromethION 2 Solo. Clinically relevant gene fusions, including BCR::ABL1 and ETV6::RUNX1, were reliably detected. We successfully implemented a transcriptome-based nanopore workflow for leukemia classification in a resource-constrained setting using a machine-learning algorithm developed at the University North Carolina. The cost-effectiveness, rapid turnaround time, and portability of the nanopore platform represent significant advantages for broader implementation in LMICs.

Biography
Mayur Parihar has a dual role, one as a clinician where he heads the Diagnostic Laboratories of Molecular Pathology and Cytogenetics at Tata Medical Center, Kolkata, which is a tertiary care cancer hospital and the other as a clinical researcher where he is in charge of the Tata Translational Cancer Research Center. Combining his two roles, Mayur’s work is focused on exploring genomic technologies and translating his research into implementation and delivery of effective, efficient, and affordable genomic solutions in clinics.

Abstract
The North Thames Genomic Laboratory Hub is one of seven hubs providing National Health Service (NHS) genomic testing, including specialist testing for 14 disease areas for half of England. The Bioinformatics team develops pipelines to interpret complex genomic data, enabling clinical scientists to identify disease-causing variants and deliver diagnoses that support personalised care, family planning, and access to targeted therapies. Long-read sequencing can improve diagnostic yield by resolving low complexity and repetitive regions, enhancing structural variant detection, and enabling phasing. However, clinical workflows and expertise are centred around short-read technologies. This project aimed to develop and integrate an Oxford Nanopore Technologies long-read sequencing pipeline into the existing clinical workflow. We developed a flexible long-read analysis framework supporting trio and singleton genome sequencing, virtual panels, targeted amplicon, and adaptive sequencing. Dedicated workflow ‘arms’ for each test type are embedded within the clinical system, with GitLab version control ensuring auditability. Automated execution via standardised sample sheets, run naming, and configuration files enable integration with existing workflow management. Analyses follow EPI2ME recommendations, and outputs feed directly into our bespoke tertiary analysis package, which offers configurable filtering modes. Results are presented within our in-house variant interpretation platform, allowing scientists to review long-read data alongside short-read outputs in a familiar interface. This integrated approach streamlines accreditation, provides control over release cycles, maintains robust data governance, and removes the need for scientists to learn new tools or work across multiple platforms. The framework establishes a scalable foundation for rapidly translating emerging long-read methods into NHS service.

Biography
Ashley Pritchard is a Health and Care Professions Council (HCPC) registered Senior Translational Bioinformatician at the NHS North Thames Genomic Laboratory Hub, with a research background and PhD in cancer genomics and over eight years of clinical experience. She is currently funded by the National Institute for Health and Care Research (NIHR) Biomedical Research Centre to bridge research and clinical service delivery by translating emerging genomic technologies into NHS practice. Her current focus is on implementing long-read sequencing technologies with demonstrated clinical utility, integrating them into routine diagnostics to enhance workflows and improve patient care.