Science unlocked: publication picks from September 2025

In this monthly series, we share a selection of recent publications in which Oxford Nanopore sequencing was used to unlock novel insights. Spanning from AI-empowered metagenomics, to single-cell research, to pathogen surveillance in wetlands, these studies showcase the advances in scientific research made possible by Oxford Nanopore sequencing.

Featured in this edition:

1. Nanopore squiggles show bacterial survival rates

2. What's lurking in wetland waters

3. Classifying acute leukaemia in under two hours

4. Single-cell genomic, epigenomic, and transcriptomic insights — in one go

5. Uncovering new variants behind hypotonia

6. Shining a light on a diagnostic blind spot

Microbiology

1. Nanopore- and AI-empowered metagenomic viability inference (GigaScience)

A new AI-driven approach uses raw Oxford Nanopore ‘squiggle’ data to distinguish viable from dead microorganisms — overcoming a key limitation of traditional metagenomic sequencing methods. Knowing whether microbes are alive matters because DNA from dead cells can persist and skew analyses, leading to false conclusions about antibiotic response, infection risk, microbial activity, and ecosystem function. By accurately classifying UV-killed and viable E. Coli DNA, this computational framework could pave the way for viability-aware metagenomics in environmental, veterinary, and clinical applications.

Key points:

• Conventional metagenomics cannot differentiate between live and dead cells

• Viability insights are vital for understanding ecosystems and pathogen virulence

• Ürel et al. present a fully computational, bias-free framework using Oxford Nanopore raw ‘squiggle’ data

• Residual Neural Network 1(ResNet1), a deep neural network, predicts microbial viability with high accuracy

• The model can generalise across taxonomic boundaries, applying the signal patterns to other bacterial species that the model wasn’t trained on, such as Chlamydia suis

Figure: sequencing read-level comparison of ResNet1 classifications of the E. coli test and the C. suis datasets. Top row: binary model predictions for viable and dead E. coli and C. suis, respectively, at the optimised prediction probability threshold of 0.5; bottom row: respective normalised distributions of model prediction probabilities across all sequencing reads. For E. coli, all 59,171 ‘viable’ and 72,207 ‘dead’ sequencing reads from the test dataset are shown; for C. suis, all 54,853 ‘viable’ and 23,073 ‘dead’ sequencing reads from both biological replicates are shown. Figure redistributed from Ürel et al. 2025 under Creative Commons Attribution License CC BY 4.0.

2. Real-time genomic pathogen, resistance, and host range characterisation from passive water sampling of wetland ecosystems (bioRxiv)

Wetlands are vital for global biodiversity but also hotspots for pathogen and antimicrobial resistance (AMR) transmission. Pathogen detection is challenging due to low microbial concentrations in water. Using Oxford Nanopore sequencing with passive water samplers, Perlas and Reska et al. successfully profiled wetland microbiomes, detected clinically relevant AMR genes, and identified potential hosts. This study demonstrates a cost-efficient, real-time approach that provides valuable insights into how the environment, wildlife, livestock, and human populations are intrinsically linked through microorganisms.

Key points:

• Wetlands support ~40% of global biodiversity and monitoring them provides early warnings of pathogen and AMR transmission

• Perlas and Reska et al. aimed to assess the potential negative implications of human activities on these environments and the risk to public health

• They found that wetlands altered by human activity had over 13-fold more pathogen-associated reads than natural wetlands

• The findings linked AMR genes to microbial hosts and offered insights into the relationship between wildlife and livestock regarding the spread of avian influenza virus

• Oxford Nanopore sequencing provides a holistic and cost-efficient approach for real-time surveillance of the wetland ecosystem

Read our pathogen metagenomics workflow overview

Cancer research

3. Rapid epigenomic classification of acute leukaemia (Nature Genetics)

Acute leukaemia is an aggressive and rapidly progressing blood cancer that requires immediate and accurate diagnosis to start the right treatment. Standard diagnostic tests for acute leukaemia can take several weeks, delaying treatment decisions. Using Oxford Nanopore sequencing with MARLIN — a machine learning method trained on methylation profiles — researchers classified leukaemia subtypes in under two hours from sample receipt. Matching standard diagnostics and revealing hidden genetic drivers, this approach demonstrates the potential for rapid, comprehensive cancer characterisation.

Key points:

• Steinicke and Benfatto et al. developed MARLIN (methylation- and AI-guided rapid leukaemia subtype inference), a neural network for classifying acute leukaemia using sparse DNA methylation data

• The model was trained on 2,540 reference methylation profiles, defining 38 distinct leukaemia subtypes from Illumina datasets

• MARLIN achieved 96.2% concordance with conventional diagnostic results, correctly classifying 25 out of 26 cases

• The approach identified cryptic genetic drivers such as DUX4 rearrangements that are often missed by standard diagnostic tests

• Beyond rapid classification, this workflow offers deeper insights into tumour heterogeneity and the molecular landscape of acute leukaemia than standard tests

Watch Salvatore Benfatto's talk at London Calling 2025

Read our case study on this research

4. Long-read single-cell genome, transcriptome and open chromatin profiling links genotype to phenotypes (bioRxiv)

Cancer is a heterogenous disease and current single-cell multiomic methods only provide limited genomic information. Here, the authors utilised single-cell Oxford Nanopore sequencing to investigate tumour evolution and why some cancer patients relapse after anti-CD19 chimeric antigen receptor (CAR) T-cell therapy. Using Oxford Nanopore sequencing, Pančíková, Cools, and Eftychiou et al. developed SPLONGGET, a custom workflow that simultaneously captures genomic, epigenomic, and transcriptomic information from individual cells. The authors applied this method to research samples isolated from a patient with leukaemia, revealing genetic changes linked to therapy resistance.

Key points:

• Most existing single-cell methods are either limited in molecular detail or low-throughput plate-based assays

• SPLONGGET retains DNA fragments of all lengths and omit the cDNA fragmentation step of the 10x Genomics Single Cell Multiome ATAC + Gene Expression assay for Oxford Nanopore sequencing

• This method enabled reliable detection of small variants, allele-specific copy number alterations, structural variants, gene expression data (full-length transcripts and gene fusions), and open chromatin patterns

• Oxford Nanopore sequencing achieved 79–93% single-cell genome coverage at ≥5x compared with just 6% from Illumina short-read data

• This research provides a versatile framework for studying tumour evolution, CAR T-cell resistance, and other complex cellular systems such as viral integration or ageing tissues

Watch Ruben Cools present this research at London Calling 2025

Find out more about unravelling tumour biology with single-cell Oxford Nanopore sequencing in this case study

Read our single-cell transcriptomics workflow overview

Human genetics

5. Long-read genomic and epigenomic profiling enhances timely comprehensive variant discovery in hypotonia and muscle weakness (Research Square)

Hypotonia (decreased muscle tone) is a key symptom in many disorders, but it has diverse genetic causes so can require multiple tests to obtain a genetic diagnosis. Short-read sequencing often fails to uncover the underlying variant as it cannot detect complex structural variants, repeat expansions, or epigenetic modifications. In this study, Abuijlan et al. used Oxford Nanopore sequencing to successfully identify all known pathogenic variants in a reference cohort, revealing the causative variants in previously unsolved cases. This single-test approach demonstrated strong potential to streamline and accelerate rare disease identification in the future.

Key points:

• Oxford Nanopore sequencing was validated on a reference-positive cohort with known diagnoses, where it identified all known pathogenic variants, before being applied to an unsolved cohort with unexplained hypotonia

• In the unsolved cohort, Oxford Nanopore data confirmed the causative variant (de novo COL6A3 deletion) in one research sample and suggested the genomic cause (aberrant methylation and copy number at POMK) in a second research sample

• Altogether, Oxford Nanopore data identified potential genomic causes of hypotonia in an additional 14% of the research samples

• For patients previously requiring sequential testing, Oxford Nanopore potentially reduced diagnostic expenses by 71.3% (saving an average of $2,478 per patient), and potentially reduced time-to-diagnosis by 85% (from 168 days down to 25 days) — the same time taken for single-test patients

• Across the whole cohort, Oxford Nanopore sequencing has the potential to reduce testing costs by 37.9% (an average saving of $611 per patient)

• Although currently approved for research use only, this work highlights the potential of Oxford Nanopore sequencing to uncover complex causative variants and streamline rare disease analysis in the future

Figure: overview of the study design and workflow. The study consisted of three main steps: (1) retrospective cohort analysis (n = 227), including clinical evaluation, hypotonia classification, standard-of-care genetic testing, diagnostic yield, and time-to-diagnosis; (2) development and validation of the Oxford Nanopore sequencing workflow (n = 15), covering both wet lab procedures and dry lab analysis with the tools used; and (3) evaluation of clinical utility (n = 14) using the optimised workflow. Figure redistributed from Abuijlan et al. 2025 under Creative Commons Attribution License CC BY 4.0.

Read our white paper on human and clinical research

6. Deep intronic SVA_E insertion identified as the most common pathogenic variant associated with Canavan disease: a diagnostic blind spot (Neurology Genetics)

Canavan disease is an autosomal recessive neurodegenerative disorder caused by biallelic pathogenic variants in the ASPA gene, leading to progressive and irreversible motor and cognitive decline. Short-read sequencing is traditionally used for genetic testing, but has often been unsuccessful for Canavan disease. The reason for this has now become clear — Canavan disease is thought to be caused by a large structural variant beyond the detection capabilities of short-read technologies. Using Oxford Nanopore sequencing, Dominguez Gonzalez and Bell et al. uncovered a retrotransposon insertion present in all eight research samples but missed by previous clinical tests. Upon searching population databases, they found that the variant exists in individuals from many different ancestries and could represent the most common pathogenic cause of Canavan disease.

Key points:

• Research samples were obtained from eight participants with a confirmed Canavan disease diagnosis based on clinical, biochemical, and neuroimaging evidence, but with unknown pathogenic variants despite short-read sequencing

• Using targeted Oxford Nanopore sequencing, the researchers identified an approximately 2,600-bp SVA_E retrotransposon intronic insertion in ASPA in all eight individuals

• Due to the repetitive nature and size of the SVA_E insertion, long reads are required to accurately characterise the genomic variant

• The frequency of this variant in population databases suggests it is the most common pathogenic variant in ASPA and is present across multiple ancestry groups

• This research demonstrates that Oxford Nanopore sequencing is capable of identifying new classes of variants for genetic disorders, particularly where standard short-read sequencing has failed

Check out our getting started guide for human genomics

Inspired? Apply Oxford Nanopore sequencing to your own research questions and you'll never see sequencing the same way again. Explore the nanopore sequencing solution.

Oxford Nanopore Technologies products are not intended for use for health assessment or to diagnose, treat, mitigate, cure, or prevent any disease or condition.