Abstract
Ribosome profiling (Ribo-seq) represents a significant advance in translatomics research. This technique enables the precise measurement of global and in vivo translation dynamics, the quantification of translation, and the identification of active translated small open reading frames (sORFs). While several databases have been developed to focus on the translatome, comprehensive databases dedicated specifically to analyses of translation and sORFs in prokaryotes remain scarce. Therefore, RiboMicrobe (https://rnainformatics.org.cn/RiboMicrobe/ and https://rnainformatics.cn/RiboMicrobe/) develops a comprehensive database tailored for Ribo-seq data from prokaryotic microorganisms, Accompanying this database, it also introduces two novel sORF prediction models based on transformer-based deep learning architecture, sORFPredRibo and sORFPred, to support translatomics analyses and sORF annotation. Currently, RiboMicrobe encompasses 891 Ribo-seq, 369 matched RNA-seq, and 62 proteome datasets from 36 prokaryotes and 2 viruses, and provides users with intuitive web interfaces to easily access and explore information of interest. In addition, a suite of bioinformatics tools encompassing three functional categories: visualization tools (Browse, JBrowse, and mRNABrowse) is developed for data exploration; predictive algorithms (sORFPred and sORFPredRibo) for sORFs prediction; and comparative analysis utilities (DiffTE, DiffCO, and BLAST) for functional investigations. It is believed that the diverse data and capabilities of RiboMicrobe will advance the field of microbial translational research substantially.