Genomics England is a government‑owned company, working in close partnership with the NHS to embed genomics in routine care at national scale and support research and innovation across the UK. Professor Scott will describe the organisation's work with Oxford Nanopore in areas spanning cancer diagnostics to discovery in rare disease and newborn sequencing.

Rapid genome sequencing is increasingly important for critically ill neonates and children with suspected genetic disorders, where a timely molecular diagnosis can directly inform treatment decisions, prognosis, surveillance, and reproductive planning. Although rapid short-read exome or genome sequencing (srGS/ES) have demonstrated significant clinical utility, they may not be able to detect or fully resolve complex genomic variations, including structural variants (SVs), repeat expansions, and epigenetic abnormalities.  Long-read genome sequencing (lrGS) offers a more comprehensive approach by generating longer sequencing reads that span clinically relevant genomic regions. Meta-analysis of 89 studies across 22 countries and regions, encompassing 43,695 individuals with rare diseases, reported a pooled diagnostic yield of 28% for combined short-read and long-read sequencing approaches (95% CI: 17-42%). Notably, among patients who remained undiagnosed after srGS, lrGS provided an additional diagnostic yield of 33% (95% CI: 14-61%). Furthermore, the negative impact of high SV burden on diagnostic yield was reduced in lrGS-only analyses, highlighting the strength of long-read technologies in resolving complex genomic variation. Despite these advances, evidence supporting the implementation of rapid lrGS in critical-care settings remains limited.  The Hong Kong Genome Institute (HKGI) has established a rapid Oxford Nanopore Technologies (ONT)-based lrGS workflow to support genetic diagnosis in critically ill patients. The workflow delivers results within 14 days and enables simultaneous genome-wide detection of single-nucleotide variants, small insertions and deletions, SVs, repeat expansions, haplotype phasing, and DNA methylation from a single assay. Representative cases demonstrate the ability of lrGS to resolve clinically significant genomic findings that are challenging to short-read approaches. These findings highlight the potential of integrating lrGS for rapid genetic diagnosis to support timely and actionable care in critically ill settings.

Exome sequencing (ES) has been widely used to identify the genetic causes of human neurological disorders ranging from childhood to adulthood; however, its diagnostic yield remains approximately 30–40%. Additional short-read genome sequencing (srGS) analyses for the remaining unsolved 60–70% of cases can identify pathogenic genetic abnormalities in approximately 21% of patients, although only about 9% harbor causative variants uniquely detectable by srGS (npj Genomic Medicine, 2025). In other words, more than half of the pathogenic variants identified by srGS could theoretically have been detected by ES. Our laboratory has been conducting long-read genome sequencing (lrGS) analyses for genetically unresolved neurological disorders. In this presentation, we will describe analytical approaches focusing on structural variants (SVs) using lrGS and discuss the extent to which these approaches contribute to solving previously undiagnosed neurological disease cases.

Aggressive blood cancers such as acute leukaemia can present with life-threatening clinical consequences. Whilst urgent therapy is required - an accurate diagnosis must first be established. Molecular characterisation of acute leukaemia increasingly underpins choice of effective frontline therapy in acute leukaemia however these approaches can often take days to weeks for results. This presentation will cover our approach to using Oxford Nanopore Technology to perform ultra-rapid molecular subclassification of acute leukaemia to improve patient outcomes for both adult and paediatric presentations.

Background: Long-read sequencing is widely seen as research-grade and difficult to translate into clinical practice, requiring bespoke DNA extraction and laboratory processes which are hard to accredit. We share our experience in translating long-read whole genome and adaptive sampling applications into a routine accredited diagnostic service. Methods: ONT long-read whole genome sequencing (lrWGS) underwent equivalency validation against accredited short-read WGS (srWGS), using gold-standard controls and prior clinical samples across multiple processing conditions and variant types. Adaptive sampling for repeat expansion disorders was assessed across 30 complex diagnostic cases and 88 archival samples using a 60+ region panel. Results: lrWGS matched or exceeded srWGS in all analysed categories, except small InDels in low-complexity regions; apparent excess false-positive SNVs localised to complex regions, alternative contigs and GRCh38 errors. Performance held on both fresh and archival DNA samples processed with routine, high-throughput sample preparation methods. Adaptive sampling for repeat expansions was >99% accurate for assignment of pathogenicity, with highly reproducible sizing. lrWGS is now clinically accredited at VCGS, promising increasing improvements for diagnostic genome reporting including phasing and better resolution of high homology regions, with adaptive sampling technologies progressing through accreditation. Conclusions: ONT sequencing is ready for clinical implementation without the need for special sample handling. Using adaptive sampling, one platform consolidates workflows including WGS, CNV, methylation, repeat sizing and any number of targeted assays. Repeat expansion, structural variant and methylation assays are progressing to accreditation, extending diagnostic reach while streamlining laboratory workflows.

Short-read sequencing has limited diagnostic utility for phasing variants, analysing genomic regions with high homology, identifying structural variants or resolving gene expansions. Nanopore long-read sequencing technology is able to overcome these limitations through the analysis of long sequences and haplotypes. Data from a validation study to implement clinical diagnostic Nanopore long read sequencing is presented with a robust, end-to-end genomic bioinformatic workflow and examples of results that provided diagnoses for patients after clinical implementation.

As the cost and accuracy of genome sequencing continue to improve, the bottleneck in many clinical applications (e.g. rare disease diagnosis and population genome screening) is the availability of expert curation capacity. Resolving this bottleneck will require increasingly automated approaches to variant prioritisation and evidence synthesis, ensuring that expert human curators can focus their attention on the most important variants, and have easy access to all of the information required to make a confident clinical decision. In this talk I will discuss our team’s work in developing and validating pipelines for automated genomic reanalysis of undiagnosed rare disease families, as well as emerging data on the value of agentic AI to streamline evidence aggregation and variant discovery. Finally, I will discuss the need for more diverse genomic reference datasets to ensure that automated and AI tools work effectively for all patients.

A fundamental challenge in human genomics is the development of reference frameworks that accurately represent population diversity and capture the full spectrum of genetic and regulatory variation. Many existing resources remain limited by ancestry bias and linear representations, constraining basic research into human evolution, population structure, and genotype–phenotype relationships. In this talk, I will describe how large-scale pangenomic initiatives in Asia are addressing these challenges by scaling from national to continental efforts. I will first present advances from the Chinese Population Pangenome Consortium (CPC) Phase II, which has generated population-representative, long-read–based reference resources across diverse Chinese populations. These data substantially improve the resolution of structural variation, complex genomic regions, and haplotype diversity, enabling refined inference of population structure, demographic history, and signals of local adaptation. By moving beyond single-reference models, CPC provides a more accurate substrate for comparative and evolutionary genomic analyses. Building on this foundation, the Asian Population Pangenome Consortium (APC) extends the pangenomic framework across Asia through coordinated international collaboration and federated data governance. APC aims to construct high-quality X-omic reference resources that integrate genome sequence with additional regulatory and functional layers, enabling systematic investigation of how genetic and regulatory variation are distributed across populations. I will discuss how scaling pangenomics from CPC to APC advances core questions in population genetics and human evolutionary biology, while establishing a rigorous and inclusive reference foundation upon which future biomedical and global health applications can be built.

The long-term hereditary consequences of ionizing radiation on the human germline remain a central question in radiation biology, yet prior studies of atomic bomb survivors have lacked the statistical power and genomic resolution to detect heritable genetic damage conclusively. Here we present a large-scale whole-genome sequencing (WGS) study of Korean atomic bomb survivor (K-ABS) families spanning three generations, including directly exposed first-generation survivors, their offspring, and grandchildren. To enable de novo mutation (DNM) detection in families lacking complete parent–offspring trios, we developed an incomplete trio pipeline that infers callable genomic regions from sibling identity-by-descent segments and applies platform-specific and cohort-level filters to minimize false positives, with calls validated by orthogonal sequencing approaches. Offspring of survivors exposed at closer distances to the hypocenter show a significantly elevated per-generation DNM burden, consistent with a radiation dose-dependent increase in germline mutation rate. Among likely damaging DNMs identified in offspring, a subset maps to genes with established disease associations, suggesting potential clinical relevance of radiation-induced germline variants. These findings provide the first WGS-scale evidence of heritable genomic instability in offspring of atomic bomb survivors and establish a broadly applicable computational framework for multi-generational DNM analysis in cohorts with incomplete familial sampling.