More than 170 chemical modifications have been identified in RNA molecules, playing crucial roles in regulating RNA metabolism and function. Unlike traditional indirect detection methods, Oxford Nanopore direct RNA sequencing enables direct RNA reading. This technology facilitates the simultaneous detection of multiple RNA modifications in a single experiment by analyzing modification-induced disturbances in the current signal and base-calling errors through computational modeling. This review systematically summarizes the current landscape of computational methods developed for RNA modification detection using this platform. It highlights two primary methodological categories—supervised machine learning and differential comparison—for identifying key modifications such as N6-methyladenosine, pseudouridine, and 5-methylcytosine, while discussing their underlying principles, characteristics, and suitable applications. Furthermore, the review outlines the technology’s potential in mapping transcriptome-wide modifications, elucidating their biological functions, and discovering disease biomarkers. Finally, we discuss existing challenges and future directions, particularly concerning the simultaneous detection of multiple modifications, improving model generalizability, and establishing standardized analytical frameworks.
LUO et al. (Fri,) studied this question.