Language-Inspired Modeling Reveals Redundant Encoding of N4-acetylcytidine(ac4C) Modifications in mRNA

Abstract

The ac4C modification on mRNA has been demonstrated to be associated with various diseases; however, its molecular mechanism remains unclear. The wet lab experiments produced relatively rough data, which lack precise ac4C modification sites, and extracting valuable information from such data remains a challenge. In this study, we integrate linguistics, traditional machine learning, and deep learning, establishing a link between the understanding of mRNA data and natural language processing (NLP). Through our analysis, we successfully revealed key information about ac4C in mRNA and uncovered the information storage mechanism of ac4C redundancy on a single sequence. This redundant information storage method in mRNA facilitates the transmission of ac4C information and promotes the enrichment of ac4C.