Periodic Reporting for period 1 - PaedsSeq (Nanopore sequencing of cell-free DNA for the sensitive detection, molecular profiling and monitoring of childhood cancers)
Berichtszeitraum: 2023-06-01 bis 2025-05-31
Paediatric cancer is the leading cause of death in children post infancy in the Western world. Solid tumours, including tumours of the central nervous system (CNS), neuroblastomas, and bone and soft tissue sarcomas, are amongst the poorest outcome cancers in children (5-year survival rates <50%). These account for >50% of cases and contribute to up to 60% of paediatric cancer deaths. Several clinical and genomic profiling studies in Europe, such as the UK Stratified Medicine Paediatrics (SMPaeds) programme, have shown that a substantial proportion of childhood malignancies contain actionable tumour alterations. This knowledge has guided the development and delivery of several successful clinical trials for targeted precision therapies. However, despite these important advances, the majority of children are still treated with multiple rounds of conventional chemotherapy, and approximately 40-60% of children with solid tumours relapse with chemo-resistant disease. At this stage, targeted therapies are the only chance of survival. Yet, only a small minority of patients (<20%) with refractory disease are currently enrolled in molecularly guided treatment regimens. This missed opportunity is the result of (1) delayed relapse detection due to the lack of feasible molecular tests to monitor disease progression; and (2) major sampling challenges associated with obtaining tissue material for genomic tumour profiling: in children, tissue biopsy requires general anaesthesia, is often not physically feasible, and commonly results in insufficient tissue for both pathology and molecular analysis.
Minimally invasive molecular profiling using liquid biopsy analyses could provide a powerful alternative to circumvent tumour sampling challenges. Liquid biopsy refers to the analysis of cell-free DNA (cfDNA) from body fluids, primarily blood, but also cerebrospinal fluid (CSF) or urine, for the detection of tumour-specific DNA aberrations. This is possible because in cancer patients cfDNA also contains small proportions of circulating tumour-derived DNA (ctDNA) released by cancer cells. Importantly, liquid biopsy analyses are highly amenable to serial sampling, thus allowing longitudinal monitoring of treatment response and disease evolution. Molecular profiling of cfDNA using DNA sequencing methods have already proven very promising for the detection of genomic aberrations in common adult malignancies as well as in some paediatric cancers. However, current liquid biopsy approaches relying on conventional sequencing are limited by (a) low sensitivity and specificity, in part owing to very low concentrations of ctDNA in the blood, (b) high upfront costs associated with current sequencing technologies; and (c) slow turnaround times. These challenges drastically limit large-scale implementation of ctDNA methods in most health-care settings across the EU. In contrast, newer long-read sequencing platforms have faster turnaround times, are more affordable, and significantly smaller than conventional sequencers, making them easy to deploy and implement in healthcare settings. In addition, real-time nanopore sequencing facilitates data analyses to be coupled to the sequencing process, which significantly reduces turnaround times from several weeks to hours. This is critical in clinical practice as long turnaround times delay therapeutic intervention. Importantly, nanopore sequencing is capable of reading individual, native DNA molecules without the need for preamplification steps. This allows simultaneous detection of alterations of the DNA sequence, including copy number aberrations (CNAs), single nucleotide variants (SNVs) and structural variants (SVs), as well as DNA modifications (i.e. DNA methylation) from the same assay. This multi-modality would be hugely beneficial for the genomic testing of paediatric cancers, and drastically increase testing sensitivity.
However, the utility of long-read sequencing for cfDNA analyses remains largely unexplored. Therefore, the overall aim of this project is to evaluate the clinical utility and feasibility of long-read sequencing for cfDNA analysis. Further, this project has the overarching aim of developing a novel multi-modal liquid biopsy test to enable highly accurate diagnosis, monitoring and early relapse detection of childhood cancers using long-read whole genome sequencing.
Minimally invasive molecular profiling using liquid biopsy analyses could provide a powerful alternative to circumvent tumour sampling challenges. Liquid biopsy refers to the analysis of cell-free DNA (cfDNA) from body fluids, primarily blood, but also cerebrospinal fluid (CSF) or urine, for the detection of tumour-specific DNA aberrations. This is possible because in cancer patients cfDNA also contains small proportions of circulating tumour-derived DNA (ctDNA) released by cancer cells. Importantly, liquid biopsy analyses are highly amenable to serial sampling, thus allowing longitudinal monitoring of treatment response and disease evolution. Molecular profiling of cfDNA using DNA sequencing methods have already proven very promising for the detection of genomic aberrations in common adult malignancies as well as in some paediatric cancers. However, current liquid biopsy approaches relying on conventional sequencing are limited by (a) low sensitivity and specificity, in part owing to very low concentrations of ctDNA in the blood, (b) high upfront costs associated with current sequencing technologies; and (c) slow turnaround times. These challenges drastically limit large-scale implementation of ctDNA methods in most health-care settings across the EU. In contrast, newer long-read sequencing platforms have faster turnaround times, are more affordable, and significantly smaller than conventional sequencers, making them easy to deploy and implement in healthcare settings. In addition, real-time nanopore sequencing facilitates data analyses to be coupled to the sequencing process, which significantly reduces turnaround times from several weeks to hours. This is critical in clinical practice as long turnaround times delay therapeutic intervention. Importantly, nanopore sequencing is capable of reading individual, native DNA molecules without the need for preamplification steps. This allows simultaneous detection of alterations of the DNA sequence, including copy number aberrations (CNAs), single nucleotide variants (SNVs) and structural variants (SVs), as well as DNA modifications (i.e. DNA methylation) from the same assay. This multi-modality would be hugely beneficial for the genomic testing of paediatric cancers, and drastically increase testing sensitivity.
However, the utility of long-read sequencing for cfDNA analyses remains largely unexplored. Therefore, the overall aim of this project is to evaluate the clinical utility and feasibility of long-read sequencing for cfDNA analysis. Further, this project has the overarching aim of developing a novel multi-modal liquid biopsy test to enable highly accurate diagnosis, monitoring and early relapse detection of childhood cancers using long-read whole genome sequencing.
During the course of this Marie Curie Fellowship, significant progress has been made towards advancing long-read sequencing technologies and computational methods for the analysis of both tumour genomes and cfDNA for the detection and characterisation of genomic alterations.
In collaboration with my secondment partner (Prof Andrew Beggs) at the University of Birmingham, we developed highly optimised methods and protocols for Nanopore sequencing of cfDNA with highly improved sequencing outputs. A patent application has been filed for these approaches.
To advance and facilitate highly robust and accurate analysis of long-read sequencing, we have also co-developed SAVANA (in collaboration with a PhD student of the host lab, the Cancer Genomics Group at the EMBL-EBI led by Dr Cortés Ciriano). SAVANA is a computational algorithm for the joint analysis somatic structural variants (SVs), somatic copy number aberrations (SCNAs), tumour purity and ploidy using long read sequencing data. Using best practices for benchmarking our algorithms in comparison to other currently existing methods across long-read sequencing data of 99 human tumour and matched normal samples, we show that SAVANA has significantly higher sensitivity and specificity, outperforming other tools.
In addition, we have developed computational tools, such as COPYBARA-cf, and approaches for the detection and multi-modal integration of tumour signal, including SCNAs, fragmentomic and epigenomic aberrations. COPYBARA-cf is the first long-read cfDNA copy number tool, and reports both the tumour burden (tumour purity) and extend of genomic abnormality detected in cfDNA samples. Overall, our work demonstrates the versatility of Nanopore sequencing to detect clinically relevant aberrations, such as ALK and MYCN amplifications, from low input cfDNA. Using matched Illumina sequencing, we show that the genomic aberrations detected using Nanopore sequencing are highly concordant. In addition, by harnessing the ability of nanopore sequencing to read out epigenetic modifications, we further developed highly accurate methylation-based cancer-type classifiers, and established methods to integrate copy number, fragmentomics and methylation signals, which facilitate disease monitoring and relapse detection in longitudinal plasma samples.
Notably, throughout this project, we have performed long-read (Nanopore) sequencing and analyses of >1000 cfDNA samples (the largest Nanopore cfDNA study to date). In doing so, we established important proof-of-principle data highlighting the clinical utility and feasibility of these approaches for cfDNA analysis and detection of tumour signal from liquid biopsies.
In addition, the highly interdisciplinary nature of this project, allowed contributions and highly successful collaborations in the wider field of cancer genomics, that further led to additional research papers and outputs.
In collaboration with my secondment partner (Prof Andrew Beggs) at the University of Birmingham, we developed highly optimised methods and protocols for Nanopore sequencing of cfDNA with highly improved sequencing outputs. A patent application has been filed for these approaches.
To advance and facilitate highly robust and accurate analysis of long-read sequencing, we have also co-developed SAVANA (in collaboration with a PhD student of the host lab, the Cancer Genomics Group at the EMBL-EBI led by Dr Cortés Ciriano). SAVANA is a computational algorithm for the joint analysis somatic structural variants (SVs), somatic copy number aberrations (SCNAs), tumour purity and ploidy using long read sequencing data. Using best practices for benchmarking our algorithms in comparison to other currently existing methods across long-read sequencing data of 99 human tumour and matched normal samples, we show that SAVANA has significantly higher sensitivity and specificity, outperforming other tools.
In addition, we have developed computational tools, such as COPYBARA-cf, and approaches for the detection and multi-modal integration of tumour signal, including SCNAs, fragmentomic and epigenomic aberrations. COPYBARA-cf is the first long-read cfDNA copy number tool, and reports both the tumour burden (tumour purity) and extend of genomic abnormality detected in cfDNA samples. Overall, our work demonstrates the versatility of Nanopore sequencing to detect clinically relevant aberrations, such as ALK and MYCN amplifications, from low input cfDNA. Using matched Illumina sequencing, we show that the genomic aberrations detected using Nanopore sequencing are highly concordant. In addition, by harnessing the ability of nanopore sequencing to read out epigenetic modifications, we further developed highly accurate methylation-based cancer-type classifiers, and established methods to integrate copy number, fragmentomics and methylation signals, which facilitate disease monitoring and relapse detection in longitudinal plasma samples.
Notably, throughout this project, we have performed long-read (Nanopore) sequencing and analyses of >1000 cfDNA samples (the largest Nanopore cfDNA study to date). In doing so, we established important proof-of-principle data highlighting the clinical utility and feasibility of these approaches for cfDNA analysis and detection of tumour signal from liquid biopsies.
In addition, the highly interdisciplinary nature of this project, allowed contributions and highly successful collaborations in the wider field of cancer genomics, that further led to additional research papers and outputs.
This project has resulted in several important advancements that go beyond the current state of the art in liquid biopsies and cfDNA analysis, significantly enhancing the sensitivity and applicability of genomic diagnostics. The main innovations/research outcomes and potential impacts are as follows:
In collaboration with the University of Birmingham, we pioneered long-read sequencing approaches specifically optimised for cfDNA analysis (the novelty and impact of these approaches are underscored by a recent patent application).
Moreover, we developed novel computational algorithms and tools that allow us to fully harness the power and advantages of long-read sequencing technologies and utilise multi-modal signal integration for highly accurate detection of tumour signal from low-input cfDNA samples. The overall output of this research also provides two highly validated and needed computational tools (SAVANA and COPYBARA-cf) for the analysis of long-read sequencing data. The application of these tools will go beyond the scope of this project as they proof highly relevant to the wider field of cancer genomics. As such, the developed tools, computational approaches, and sequencing protocols resulting from this project provide powerful new resources for researchers exploring genomic alterations in both cfDNA and tumour tissues from long-read sequencing data, and will further drive the discovery of novel cancer biology insights.
Notably, this project provides critical proof-of-principle data to establish and illustrate the high feasibility and utility of long-read sequencing for cfDNA analysis, particularly in children with cancer. A second phase of the SMPaeds study is now underway, to further build and expand on this work, and implement the above developed approaches into longitudinal disease monitoring of paediatric cancer patients.
Together, these advancements significantly enhance the analytical capabilities for cfDNA research and diagnostics, positioning these novel tools and methodologies at the forefront of precision medicine.
In collaboration with the University of Birmingham, we pioneered long-read sequencing approaches specifically optimised for cfDNA analysis (the novelty and impact of these approaches are underscored by a recent patent application).
Moreover, we developed novel computational algorithms and tools that allow us to fully harness the power and advantages of long-read sequencing technologies and utilise multi-modal signal integration for highly accurate detection of tumour signal from low-input cfDNA samples. The overall output of this research also provides two highly validated and needed computational tools (SAVANA and COPYBARA-cf) for the analysis of long-read sequencing data. The application of these tools will go beyond the scope of this project as they proof highly relevant to the wider field of cancer genomics. As such, the developed tools, computational approaches, and sequencing protocols resulting from this project provide powerful new resources for researchers exploring genomic alterations in both cfDNA and tumour tissues from long-read sequencing data, and will further drive the discovery of novel cancer biology insights.
Notably, this project provides critical proof-of-principle data to establish and illustrate the high feasibility and utility of long-read sequencing for cfDNA analysis, particularly in children with cancer. A second phase of the SMPaeds study is now underway, to further build and expand on this work, and implement the above developed approaches into longitudinal disease monitoring of paediatric cancer patients.
Together, these advancements significantly enhance the analytical capabilities for cfDNA research and diagnostics, positioning these novel tools and methodologies at the forefront of precision medicine.