Tonight I watched Jeroen de Ridder from the University Medical Center Ultrecht in the Netherlands present at London Calling 2019. They spoke about “Cyclomics: ultra-sensitive nanopore sequencing of cell-free tumor DNA.” Ridder spoke about the importance of detecting the recurrence of cancer in patients. Common diagnostics may not be suitable, and liquid biopsies of the blood can help identify tumor mutations and biomarkers. Cell-free DNA (cfDNA) is in low abundance, and therefore, sensitive techniques are essential. Ridder and team developed CyclomicsSeq to capture and circularize molecules to amplify and sequence targets. The protocol is not only sensitive but also flexible and fast. Sequences have repeating units of backbone and target. Sequencing the backbone can help identify errors. The team created synthetic P53 mutants and cloned into the backbone for transfection. Pools of mutants and wild type were used to calculate the ability to identify mutations. The results were used to support the application of this approach to head and neck cancer patient samples. Using High Accuracy base calling models (HAC) reduced error rates. Interestingly, there were differences in errors and accuracy between R9.4 and R10.4 flow cells. Ridder and the team are conducting a clinical trial analyzing samples from head and neck cancer patients. As the basecallers improve, so does Cyclomics output, according to Ridder. The group is using deep learning models and training data. A spinout company called Cyclomics aims to use this approach for cancer patients. The approach is being modified for other targets and base modifications.
