Unravelling tumour biology with single-cell Oxford Nanopore sequencing

Comprehensive molecular testing for cancer diagnosis is well established, but we need a better understanding of tumour biology to improve our treatment efficacy due to ageing populations, where there is an increasing burden of haematological malignancies¹. To tackle this, researchers are turning to single-cell sequencing to unravel the complexity and treat blood cancers effectively — reducing the burden of disease.

In this case study, we showcase how researchers are performing single-cell sequencing with Oxford Nanopore technology to learn more about cancer biology, along with their preliminary findings.

Creating a full picture of immune cells

Dr Ivo Gut (National Centre for Genomic Analysis, Spain) is investigating chronic lymphocytic leukaemia (CLL), specifically to understand how CLL develops from healthy B cells. Using short-read technology, the team prepared healthy research samples from tonsils using a 10x Genomics method to perform single-cell sequencing. They annotated 121 cell types and states using multiple genomic assays, creating an atlas of healthy immune cells from tonsil samples to map the cell states and types in the fully functioning immune system².

However, Ivo found that he could generate the same data with Oxford Nanopore sequencing and access more: full-length transcripts. Ivo and his team re-sequenced the same 10x Genomics libraries prepared from healthy immune cells and also sequenced cells from CLL research samples. In this second atlas of immune cells, they generated on average 100 million Oxford Nanopore reads of 1–2 kb per sample, meaning they could ‘really cover the full length of a transcript’, in this second immune cell atlas.

From this whole-transcriptome sequencing data, Ivo could identify which genes have specific somatic mutations in CLL cells and finally ask, ‘how does [a] mutation impact the transcript profile of that particular cell?’. However, the expression levels of somatic mutations are low, as they typically only affect a few cells. Therefore, Ivo also performed targeted nanopore sequencing on the CLL research samples to achieve high-depth coverage of the genes of interest.

For one target gene, BLNK, which encodes a protein critical for normal B cell development³, the nanopore data revealed the cell types that expressed aberrant BLNK isoforms. Ivo also investigated a somatic mutation in the protein transport gene XPO1⁴, and identified approximately 50 cells from a single CLL research sample with the mutation. The nanopore data revealed that cells with the mutation also exhibited differential expression of IGKJ5, an immunoglobulin gene that may have a role in CLL⁵.

Ivo presented these preliminary results at London Calling 2025, demonstrating the additional data Oxford Nanopore sequencing reveals, which was previously inaccessible with short-read data. Ivo and his team are still exploring the nanopore data, but with access to full-length transcripts, they are advancing their research into how healthy B lymphocytes mutate and develop into CLL.

Tracking cancer cell evolution

Researchers have also been using single-cell Oxford Nanopore sequencing to investigate tumour cell evolution. Ruben Cools, from the Integrative Cancer Genomics Lab, Belgium, presented preliminary data at London Calling 2025 demonstrating how paediatric B cell acute lymphoblastic leukaemia (B-ALL) cells evolve during chemotherapy and chimeric antigen receptor T cell (CAR-T) therapy.

To capture all variants and track tumour cell evolution on a single platform, the research team developed SPLONGGET. This method utilises 10x libraries and nanopore sequencing to generate high-output multiomic data, including whole-genome, open-chromatin, and full-length transcriptome data across thousands of cells. They applied this method to the research samples isolated from a patient with B-ALL. From this individual case, they sequenced four samples from across multiple time points, starting from initial diagnosis through to the third disease relapse. Across these four time points, they could see how the tumour cells evolved and responded to treatment.

A huge amount of data was generated, enabling Ruben to perform multiple analyses investigating the different cell types and states for tumour and normal cells. Utilising the single-cell transcriptome profile, he presented a uniform manifold approximation and projection (UMAP) graph of all cell types across the four different time points, showing four distinct tumour clusters at each period that illustrate the cell evolution.

He also investigated the open-chromatin data and could see that the tumour cells had open chromatin in the ERG gene, a proto-oncogene that enhances tumour growth⁶, and increased expression of the gene regulator TCF4. Ruben explained this was significant because ‘we see that this is truly a cancer-specific regulatory programme as only these [tumour] cells have, actually, expression of ERG and TCF4.’

Ruben’s research aimed to investigate why the tumour cells became resistant to the patient’s final treatment: CAR-T therapy. This is a form of immunotherapy that is developed from the patient’s T cells to attack cancerous cells⁷. In this case, the therapy targeted the CD19 antigen, a B cell-specific marker expressed in B-ALL⁸. Using both the DNA and RNA data, he compared the B-ALL cells pre- and post-CAR-T therapy. In the DNA data, he found two mutations in cells post-CAR-T therapy, both causing intron retention as demonstrated by the RNA data, ‘hinting at a potential mechanism [as to] why these cells became resistant’.

Taking advantage of the long-read data, Ruben located the CAR-T cells in the research samples. These cells are characterised by the gene fusion of TNFRSF9-CD247 in the T cell membrane, enabling Ruben to trace the fusion back, leading him to cytotoxic T cells on his UMAP graph. Specifically, he was able to see that the cytotoxic CAR-T cells were also present in the tumour cell clusters, suggesting an interaction between the cells.

By sequencing 10x libraries with Oxford Nanopore technology, Ruben accessed a wealth of information to reveal more tumour data than before. He was also able to match cDNA and DNA 10x barcodes for multi-modal integration and created pseudobulk data for matched tumour-normal whole-genome analysis to ‘call even the most complex variation’. Here, Ruben demonstrated that with access to all genomic layers of a tumour, more tumour biology can be unravelled to help researchers understand why some treatments fail and how to improve them in the future.