Direct detection of modified DNA bases

The human genome is built out of 4 nucleotide bases A, C, T en G. Changes in the sequence by which these nucleotides occur are called mutations and these mutations, when occurring in essential parts of the genome cause disease including cancer. Recently it has become apparent that nucleotide bases can be directly modified by many chemical reactions that are induced by the cellular environment. However, detecting these modified bases in endogenous DNA has proven difficult. Cyclomics BV. together with the Center of Molecular Medicine at the UMCUtrecht have collaborated in a project to develop a new method for direct assessment of modified bases through a combination of 3rd generation sequencing and machine learning approaches. The project yielded a new way to detect oxidated G-DNA-bases (8-oxo-G) using handheld Nanopore sequencing devices. To achieve this, the team developed machine learning methods that can identify minute changes in the measurement data produced by Nanopore sequencers indicative of DNA modifications. To train such models, vast amounts of synthetic DNA molecules with predefined 8-oxo-dG modifications were created. Utilizing the method on human genome sequencing data, we demonstrated exceptionally high accuracy and were able to simultaneously measure methylation (an important DNA modification involved in gene regulation) and 8-oxo-dG at single molecule resolution. Our results pave the way to detect a wide range of synthesizable DNA base modification for future studies into modifications of the human genome. Our work has potential applications in detection of biomarkers in cell free DNA obtained by non-invasive liquid biopsies of e.g. the blood or urine, and therefore can have profound benefits for cancer diagnostics.