Oxford Nanopore User Group Meeting, Brisbane

Reference genomes have traditionally come from pure isolate cultures, yet most microbial diversity resists cultivation. Metagenome-assembled genomes (MAGs) have filled this gap but at the cost of fragmentation. Additionally, short-read assembly and binning routinely miss or mis-bin ribosomal proteins, mobile genetic elements, biosynthetic gene clusters, and antimicrobial resistance genes. Deep long-read metagenomics bypasses this trade-off: genomes reconstructed directly from complex communities can rival, and often surpass, the contiguity and completeness of existing isolate-derived references. This talk presents two case studies demonstrating the approach. In the first, samples from the gut microbiome of 51 urban pet dogs in Shanghai (sequenced to at least 20 Gbp of Nanopore plus 20 Gbp of Illumina data) yielded 2,676 MAGs spanning 320 bacterial species, roughly 72% of them near-finished. Many improved on the corresponding public reference genome for the same species. The catalog proved broadly representative of pet dogs worldwide (>90% median read mapping to external cohorts), and crucially recovered 185 circular extrachromosomal elements, including plasmids carrying antimicrobial resistance genes that also circulate in dog gut datasets globally. Living environment, not geography, emerged as the dominant structuring factor. We were also able to observe one interesting instance of structural genome rearrangements with a Fusobacterium sp. alternatively presenting as having two chromosomes or a single, fused, one. The second case study applies the same approach to 58 urban soil samples from two Chinese cities. Here, long-read assembly recovered 7,949 medium- and high-quality MAGs representing 4,171 species-level bins, over 97% previously undescribed. The contiguity unlocked secondary metabolism >30,000 biosynthetic gene clusters, substantially more complete than short-read equivalents, alongside over 2 million small protein families enriched near defense systems and mobile elements, and an extensive repertoire of latent antimicrobial resistance genes.

Epigenetic clocks based on DNA methylation patterns have emerged as powerful biomarkers of biological age, health status, and physiological resilience across humans and livestock species. However, widespread adoption of epigenetic clock technologies has been constrained by the high cost and limited scalability of conventional methylation profiling approaches, including bisulfite sequencing and array-based assays. Recent advances in Oxford Nanopore Technologies (ONT) sequencing provide an opportunity to directly detect native DNA methylation from long-read sequencing data without chemical conversion or amplification, potentially enabling rapid and cost-effective epigenetic profiling at scale.

In this study, we evaluate the feasibility of using low-coverage ONT sequencing data for accurate epigenetic age prediction. Native methylation calls were generated from nanopore signal data and integrated into prediction pipelines to construct epigenetic clock models. We investigated the trade-offs between sequencing depth, methylation site density, predictive accuracy, and per-sample cost to identify practical operating points for scalable deployment. Particular emphasis was placed on sparse methylome approaches, assessing whether robust biological age estimates can be achieved using substantially reduced sequencing coverage compared with traditional methods.

Our results demonstrate that ONT-derived methylation profiles retain strong predictive power for biological age, even under low-pass sequencing conditions. Cost modelling indicates that nanopore-based workflows could substantially reduce the per-sample cost of epigenetic clocks while simplifying laboratory protocols and improving turnaround times.

These findings support nanopore sequencing as a promising platform for scalable epigenetic biomarker development. More broadly, this work highlights the potential for native long-read epigenomics to democratise access to biological ageing assays and enable large-scale longitudinal studies that were previously cost prohibitive.

Oxford Nanopore Technologies (ONT) sequencing provides the resolution required for complex microbial genomics, yet optimal data processing and assembly of difficult architectures remain challenging. This presentation spans recent efforts to address biological complexity and bioinformatic workflow optimisation. The primary focus is on recent work characterising Enterococcus faecium isolates resistant to last-resort antibiotics. Complete assembly of these genomes revealed that key resistance determinants reside on a linear plasmid. I will outline the specific challenges associated with assembling these linear architectures and how ONT long reads were instrumental in their resolution. Additionally, I will present preliminary results from two ongoing bioinformatic projects aimed at refining ONT pipelines. The first investigates read-length-based selection strategies to improve plasmid assembly, specifically targeting the retention of small plasmids frequently lost during standard assembly processes. The second systematically benchmarks quality control (QC) tools to determine their tangible impact on downstream applications. Early results challenge standard assumptions, indicating that steps such as adapter trimming may have no meaningful impact on variant calling accuracy. These projects highlight how refining assembly approaches and interrogating standard QC assumptions can maximise the use of ONT data for AMR surveillance and plasmid biology

Methylartist is an integrated suite of tools for working with modified bases including but not limited to 5mC. Since first publishing methylartist in 2022, we have steadily added new features as it has seen increased adoption and application across the ONT community. In this talk I will run through some of our use cases in scaling methylartist to study haplotype-aware methylation in larger cohorts, as well as new features including variant-specific analysis and support for comparison with methylation calls based on C/T substitution assays.