The expanding field of epitranscriptomics might rival the epigenome in the diversity of biological processes impacted. However, the identification of multiple modification types in individual RNA molecules remains challenging. We present CHEUI, a new method that detects N6-methyladenosine (m6A) and 5-methylcytidine (m5C) in individual transcript molecules in a single condition as well as differential methylation between two conditions, using nanopore signals. CHEUI processes observed and expected signals with convolutional neural networks to achieve high single-molecule accuracy and outperform other methods in detecting m6A and m5C sites and quantifying their stoichiometry. Moreover, CHEUI’s unique capability to identify different modifications in the same signal data reveals a non-random co-occurrence of m6A and m5C in transcripts in human cell lines and during mouse embryonic brain development. CHEUI unlocks the capability of studying links between multiple RNA modifications and phenotypes, enabling the discovery of new epitranscriptome functions. Furthermore, CHEUI's training and testing protocols are adaptable to other modifications, making it a versatile RNA technology.
Overall design: We performed RNA bisulfite treatment (bsRNA-seq) following the protocol from Johnson et al. (Johnson, Z., Xu, X., Pacholec, C. & Xie, H. Systematic evaluation of parameters in RNA bisulfite sequencing data generation and analysis. NAR genomics Bioinforma. 4, lqac045 (2022). There were no deviations done for the protocol except the fact that a GE50 spin column was used for the removal of the excess of bisulfite reagent (sodium bisulfide and hydroquinone) instead of a GE25. This protocol was applied to RNA fully modified with m5C or non-modified obtained from in-vitro transcripts (IVTs) m1, m2, m3, m4 built from four non-overlapping fragments from the mouse canonical pre-rRNA (~13kb long). Sequencing of the bisulfite-treated samples was performed with Illumina.
Less...