Methods to Develop an 8-oxoG Base Modification Classifier for Direct-DNA Nanopore Sequencing

Abstract

Space exploration encounters a significant hurdle in radiation-induced oxidative damage, notably the accumulation of 8-oxoguanine (8-oxoG), associated with aging-related health concerns and cancers. Current 8-oxoG detection methods are slow or lack single-nucleotide resolution. Our aim, using Oxford nanopore technology (ONT) to classify bases based on their electrokinetic properties, is to establish a dataset for training a classifier to detect 8-oxoG in all sequence contexts. Comparing ONT cDNA sequencing with Illumina, we found moderately similar gene ontology and highly similar enriched phenotype term lists. Despite challenges in generating dsDNA libraries, we redesigned the schematic for a fully double-stranded DNA library to improve 8-oxoG classifier training. We discuss the resulting dataset and methods to enhance its quality for training a base modification model. Additionally, we explore using this protocol as a spike-in control to evaluate the accuracy of modified basecalling models.