WYMM Tour: Melbourne

Effective management of liver cancer requires regular assessment of disease progression. However, standard imaging techniques, such as MRI and contrast-enhanced CT, are costly and resource-intensive, limiting their feasibility for frequent monitoring. These methods are often inaccessible in remote regions and may not reliably detect small tumours under 2 cm, posing challenges for timely evaluation of disease progression.

We are developing a methylation-based circulating tumour DNA (ctDNA) assay for liver cancer research, designed to detect residual disease and recurrence in patients after curative treatment. This approach employs Nanopore sequencing of plasma cell-free DNA (cfDNA) samples, coupled with a single-molecule deconvolution method developed by our team. The workflow enables the detection and quantification of liver cancer-associated ctDNA with minimal sample input, offering a cost-effective and rapid alternative suitable for diverse settings.

To date, we have analysed 58 research samples from a Western Australian cohort. Future efforts will focus on incorporating a larger number of control samples from individuals with chronic liver disease but no liver cancer, to refine the thresholds for deconvolution and evaluate the sensitivity and specificity of the method. Additionally, we aim to conduct a prospective observational study to explore the potential of ctDNA results to indicate early recurrence risk following curative surgery.

Nanopore sequencing technologies (Oxford Nanopore Technologies) offer the ability to sequence longer reads, providing unique advantages for detecting complex structural variants (SVs) that are often missed by short-read sequencing. Simultaneous whole-genome sequencing and methylation profiling further strengthen its application in cancer research. In this presentation, I will highlight the development of novel pipelines and in-house tools to improve SV detection and methylation analysis. These advancements have enabled the resolution of SVs previously undetected with short-read technologies and the identification of methylation signatures. Furthermore, our findings reveal a correlation between methylation patterns and SV signatures, highlighting the potential interplay between genomic and epigenomic factors in cancer. The integration of ONT technology opens new avenues for understanding tumour heterogeneity and identifying actionable biomarkers, advancing precision oncology.

Facioscapulohumeral muscular dystrophy (FSHD) is a genetic disorder caused by epigenetic reactivation of the D4Z4 macrosatellite on chromosome 4. Variability in disease onset and severity remains poorly understood, with potential genetic and epigenetic contributions. Until recently, the D4Z4 region was inaccessible due to its size (5-100+kb) and repetitive nature (3.3kb repeat unit). The progress in long-read genomics, such as ultra-long reads and telomere-to-telomere genomes, finally enables us to fully sequence this genomic region.

We employ both whole-genome and targeted sequencing strategies to capture the full spectrum of genetic and epigenetic variation at D4Z4, related macrosatellites and disease modifier genes. A custom analysis pipeline allows us to resolve assembly issues and generate complete (epi)haplotypes in FSHD patients. We find striking methylation patterns, describe new alleles and cis-duplication events, and reveal how pathogenic variants in the epigenetic machinery affect D4Z4 regulation.

Our workflow has the potential to greatly speed up and simplify FSHD diagnosis, while providing much more information about the complete genetic make-up and epigenetic mechanisms at D4Z4. It provides a more comprehensive understanding of FSHD pathogenesis and guides the development of still-elusive treatments.

ROBIN (Rapid nanopOre Brain intraoperatIve classificatioN) is a groundbreaking tool that utilizes PromethION nanopore sequencing technology for rapid and comprehensive molecular profiling of CNS tumours. Advances in genomic sequencing have revolutionized the diagnostic pathways for brain tumours, shifting focus to epigenetic signatures for classification. ROBIN capitalizes on this shift, offering real-time, intraoperative methylome classification alongside next-day comprehensive profiling within a single assay. This approach enhances classification precision and aids surgeons in tailoring surgical strategies based on immediate genetic insights, thus balancing the risks and benefits of intervention.

ROBIN integrates dynamic adaptive sampling through tools like ReadFish, enabling targeted sequencing of critical genomic regions while also dynamically detecting deletions and insertions during sequencing. This capability allows ROBIN to prioritize and enrich sequencing data from regions of interest, capturing clinically significant structural variations in real time. By optimizing sequencing efficiency and increasing coverage of relevant areas, ROBIN reduces turnaround time and cost without compromising classification accuracy.

The tool employs a multi-classifier system—incorporating algorithms such as Sturgeon, CrossNN, and RapidCNS2—ensuring high precision in methylation-based tumour classification. Classifier performance has been validated on 50 prospective intraoperative cases, achieving robust tumour classifications within minutes and a profiling turnaround time of less than 2 hours, with 90% concordance with final integrated diagnoses.

Beyond methylation profiling, ROBIN detects single nucleotide variants (SNVs), copy number variants (CNVs), structural variants (SVs), and now deletions and insertions in real time. Its ability to provide a complete integrated profiling within 24 hours significantly enhances classification fidelity, reducing delays associated with additional assays.

ROBIN’s integration of real-time analysis, dynamic adaptive sampling, and cutting-edge algorithms represents a transformative advancement in brain tumour classification, delivering faster, more precise, and clinically actionable results to improve patient outcomes.

High-throughput sequencing has revolutionized how we think about transcriptional regulation by revealing enormous diversity in mRNA transcript expression across cell types and tissues. Emerging data has revealed that splicing is pathologically altered in cancer, including recurrent mutations in the splicing machinery, and ~30% more alternative splicing events have been observed in tumors compared to normal tissues. Notably, in some cancer patient cohorts including melanoma, the extent of transcript diversity is prognostic, and altered RNA processing has been implicated as a major driver of acquired resistance to anti-cancer therapies. We therefore hypothesise that disruptions in splicing and altered transcript diversity might help melanoma cells become more aggressive and resist current therapies. To explore this further, we have applied Nanopore long read sequencing to our preclinical mouse model of melanoma to comprehensively profile how the diversity of mRNA transcripts changes during the response to treatment. Notably, we identified thousands of “novel” transcripts, including hundreds of significant isoform-switching events, in melanoma tumours treated with targeted therapy. By using long read data as a reference transcriptome, our preliminary analysis has detected ~50% of novel isoforms in short-read data obtained from melanoma patients, indicating these novel transcripts may be clinically relevant. We have now applied this approach to a large cohort of primary and early metastatic melanoma patient samples to better understand how the diversity of mRNA transcripts changes during progression. These analyses provide new insights into the post-transcriptional landscape of metastasis and therapy resistance in melanoma and may identify new targets for mRNA-based therapies.

Collectively, rare diseases affect approximately 8% of the population and are one of the leading causes of death for children under 5. While nearly 80% are genetic in origin, a genetic diagnosis is frequently unable to be resolved. Two key reasons for this are that the genetic causes lie in parts of the genome that cannot be resolved by the short read sequencing technology that has traditionally been used, or that the genetic architecture of associated variation is too complex to easily resolve using short read methods. Long Read sequencing, offers a potential pathway to address these challenges.

The Victorian Clinical Genetic Services (VCGS) are trialling use of Oxford Nanopore Technologies Long Read Sequencing (ONT LRS) as a method for improved diagnosis and gene discovery for rare disease patients. To evaluate the utility of ONT LRS, VCGS sequenced a series of case studies containing pathogenic genetic variation that have been traditionally difficult to resolve via short read methods. Subsequently, we have evaluated use of adaptive sampling for targeted enrichment of loci of interest and begun the process of comprehensive benchmarking required for clinical validation of this technology as an option for clinical testing.

Our findings demonstrate that as long as certain specific limitations are accounted for, ONT LRS is well within an accuracy range supportable for clinical testing. We note specific circumstances where ONT LRS can be confounded by complex variation, and additional interpretation support and methods development may be required to accurately resolve the nature of re-arrangements. Despite this, we consider that ONT LRS is highly effective at resolving the large majority of structural variants, many of which are difficult or impossible to accurately characterise using short read methods.

Infections caused by Gram-negative organisms resistant to third-generation cephalosporins and carbapenems are a major global health problem with few treatment options. Resistance is mediated by extended-spectrum beta-lactamases (ESBLs) or carbapenemases that are frequently located on large plasmids capable of transfer between different bacterial species. Detecting plasmid transmission in hospitals remains difficult with no gold-standard methods. We therefore aimed to develop a novel methodology for detecting plasmid transmission using Oxford Nanopore long-read whole genome sequencing.

We conducted a 6-month prospective study in the haematology/bone marrow transplant ward of our hospital to detect plasmid transmission between patients. Patients were screened for colonisation on admission, weekly, then on discharge. All clinical isolates causing infection were collected.

We collected 171 ESBL and/or carbapenemase isolates from 253 patients. Of these 171 isolates, only 11 came from clinical infections, the remaining 160 were colonisation. Isolates were long-read sequenced and assembled with Hybracter. Antimicrobial resistance genes were detected with AMRFinderPlus. Most isolates were E. coli (n=89), followed by Klebsiella (n=42), Enterobacter (n=30) and Citrobacter (n=7). Acquired ESBL/carbapenemase genes were found on plasmids in 68% of genomes.

We extracted 107 ESBL/carbapenemase plasmids and applied a systematic plasmid comparison protocol, which combines plasmid replicon information, plasmid distances, single nucleotide variants, and patient movement data. This revealed that usage of a single tool was often insufficient when attempting to tease apart complex plasmid communities. We found plasmid transmission to be rare in our dataset, but strain transmission was frequent. This study demonstrated that there is significant AMR transmission occurring between patients on this ward, which can then lead to infections by these resistant colonising isolates.

Clinical metagenomics using nanopore sequencing has potential to transform routine microbiology practice given the ability to provide comprehensive information on all relevant pathogens in clinical samples in about 6 hours. This includes strain type and carriage of virulence or antimicrobial resistance determinants that together can inform antimicrobial treatment, infection control and public health interventions. Furthermore, given the agnostic nature of metagenomics, that same sequence dataset can in theory be used for surveillance of novel and emerging pathogens to provide a vital Biosecurity function.

It is this breadth of benefits from a single sequencing output provided in real time that holds promise of transforming microbiology practice. The potential to bring clinicians, infection control and public health professionals together in the same initial timeframe to implement co-ordinated rapid interventions.

To realise these benefits, we have embarked on a translational research journey taking respiratory metagenomics through a research phase into early clinical evaluation for patients with severe respiratory infection on an intensive care and ECMO unit. Initial studies provided proof-of-concept evidence of benefits including optimising initial antibiotic treatment in almost half of patients, as well as identifying unusual and hidden pathogens with wider infection control or public health benefit in about 10% of cases. These findings led to funding to establish an NHS England Network-of Excellence at 10 hospital sites, with respiratory metagenomic data being transferred from hospital sites to UKHSA for real-time pathogen surveillance. Based on continued progress at these pilot sites this programme has been extended across the UK. This talk will present the journey from early research through to network implementation, highlighting the progress, key findings and future requirements to establish a sustainable dua-use national pathogen surveillance system embedded in the clinical microbiology laboratory.

The What You're Missing Matters Tour showcases some of the most cutting edge genomics in Australia and abroad. This work is underpinned by continual technology developments in machine learning, chemistry, protein engineering and analysis. This talk will touch on some of the technology updates from Oxford Nanopore over the last year, and provide a glimpse forward to what is to come.