Nanopore sequencing improves the characterisation of mutations driving blood cancer

With recent advances in drug development, especially in the immunotherapy field, it has never been more important to fully elucidate the biological hallmarks and mechanisms of cancer. Scientists are deploying many different tools to answer this need. In two papers, research teams described new techniques to characterise blood cancer samples faster and more comprehensively.

Rapid analysis

Lead author Cecilia Yeung and colleagues at the University of Washington, USA, reported results of a proof-of-concept study to test the potential of nanopore sequencing for rapid identification of cancer-driving fusions in acute leukaemia samples¹. Currently, reliance on time-consuming conventional techniques such as cytogenetics or fluorescence in situ hybridisation results in long waits for diagnosis and, in many cases, the disease cannot be pinpointed completely. This new study demonstrates the future potential for generating this information faster in a clinical setting, which could make it possible to match people to the therapies most likely to be effective for their cancer and improve outcomes through rapid delivery of treatment.

A key target of this project was to identify the recurrent fusions common to leukaemia cases because they can inform risk stratification and treatment selection. The team noted that ‘Clinically defining or significant fusion genes are present in approximately 30% of acute myeloid leukemias and 30% of acute lymphoblastic leukemias’. These long elements can be difficult to detect with short-read sequencing because the fusions extend well past any individual read and may appear to align with the original genes that formed them.

Yeung et al. designed an ultra-rapid assay based on nanopore sequencing that uses Cas9 targeted sequencing of a panel of fusion genes. Thanks to real-time analysis in the cloud, analytical steps were performed while the MinION sequencing device continued running. They ran the workflow for three clinical research samples, each representing a different type of leukaemia.

In this study, nanopore sequencing allowed the team to capture timed data snippets to facilitate near-real-time analysis. Confirmation of a fusion was accepted after three calls of that fusion were found. ‘Total data analysis time to confirm if a … sample has a fusion was ~271 seconds from basecalling, alignment, fusion detection, and visualization of fusion calls’, the authors reported. They highlighted that ‘this enables a sample to fusion reporting workflow that takes an average of 5.85 hours’. These turnaround times are ‘substantially faster than any current fusion detection assays used in clinical laboratories’.

The proof-of-concept study successfully showed the potential utility of this approach for rapid fusion identification in leukaemia samples.

RNA splicing

In a separate project, Federico Gaiti and collaborators at the New York Genome Center, USA, and other institutions, described a technique to deeply characterise individual cells for more accurate differentiation of mutant and wild type cells as well as the identification of key RNA splicing factors². One goal was to characterise RNA splicing factors, which are commonly dysregulated in haematologic malignancies such as blood cancers.

Using traditional short-read RNA-seq technology, the authors noted, can result in ‘sparse and biased coverage of splice junctions’ in single-cell RNA-seq data. To address that issue, they developed GoT-Splice, based on the use of long nanopore sequencing reads, for genotyping of single-cell transcriptomes combined with CITE-seq and other molecular profiling tools.

In this project, the scientists tested their new pipeline on three bone marrow progenitor clinical research samples from myelodysplastic syndrome cases with specific mutations in the RNA splicing factor SF3B1. ‘This allowed for the simultaneous single-cell profiling of gene expression, cell surface protein markers, somatic mutation status, and RNA splicing’, they wrote.

The team chose long nanopore sequencing reads to capture full isoforms for analysis of RNA splicing events across the entire transcriptome, sequencing full-length cDNA on GridION or PromethION devices. They also added a biotin enrichment step to the process to improve efficiency by selectively amplifying full-length transcripts. Combined with a custom-built analytical pipeline, the method produced high-resolution, full-length profiles of single-cell transcriptomes. The scientists looked for intron-centric junctions to identify mis-splicing events. ‘As anticipated, we observed a 4-fold increase in the number of junctions per cell detected using full-length long-read sequencing over short-read, despite lower absolute number of [unique molecular identifiers]/cell’, they reported. Furthermore, they noted that ‘GoT-Splice afforded greater coverage uniformity across the entire transcript, compared to 3’-biased coverage in short-read sequencing’.

A deep dive into the data showed that alternative 3’ splice sites were the most common mis-splicing events. ‘Long-read sequencing also allowed us to quantify the presence of different splicing events across the same mRNA transcript’, the team wrote. They identified 428 genes that had multiple aberrant 3’ splice site events. ‘Interestingly, these cryptic 3’ splicing events tend to appear in different copies of the transcript, highlighting the unique advantages of long-read sequencing in this context’, the authors added.

The team’s analysis showed important differences between wild type and mutant SF3B1 RNA splicing factors, including mis-splicing events specific to certain cell types and cell stages. The data indicate that these dysregulation patterns could trigger the cascade of changes responsible for the phenotype. ‘Importantly, the integration of GoT with full-length isoform mapping via long-read sequencing showed that SF3B1 mutations exert cell-type specific mis-splicing, already apparent in [clonal hematopoiesis] long before disease onset’ they wrote.

The authors concluded that the GoT-Splice pipeline represents ‘an expanded multi-omics single-cell toolkit to define the cell-type specific impact of somatic mutations on RNA splicing, from the earliest phases of clonal outgrowths to overt neoplasia, directly in human samples’.