Single-molecule detecting and phasing of A-to-I RNA editing events using nanopore direct RNA sequencing

Single-molecule detection and phasing of A-to-I RNA editing events remain unresolved problems. Compared to the traditional short-read sequencing technology, the long-read and PCR-free nanopore native RNA sequencing offers great opportunity for direct RNA editing detection. Here we developed a neural network model, DeepEdit, that not only recognizes A-to-I editing events in single reads of Oxford Nanopore direct RNA sequencing, but also resolves the phasing of RNA editing events on transcripts. We illustrated the robustness of DeepEdit by applying it to Schizosaccharomyces pombe and Homo sapiens transcriptome data. We anticipate DeepEdit to be an accurate and powerful tool for both biofunction and biomedical study of RNA editing in a new perspective.