Exploring the Roles of Ribosomal Peptides in Prokaryote-Phage Interactions through Deep Learning-Enabled Metagenome Mining

Abstract Background Microbial secondary metabolites play a crucial role in the intricate interactions within the natural environment. Amongst these metabolites, ribosomally synthesized and post-translationally modified peptides (RiPPs) are becoming a promising source of therapeutic agents due to their structural diversity and functional versatility. However, their biosynthetic capacity and ecological functions remain largely underexplored. Results Here, we aim to explore the biosynthetic profile of RiPPs and their potential roles in the interactions between microbes and viruses in the ocean, which encompasses a vast diversity of unique biomes that are rich in interactions and remain chemically underexplored. We first develop TrRiPP to identify RiPPs from ocean metagenomes, a deep learning method that detects RiPP precursors in a hallmark gene-independent manner to overcome the limitations of classic methods in processing highly fragmented metagenomic data. Applying this method to metagenomes from the global ocean microbiome, we uncover a diverse array of previously uncharacterized RiPP families with great novelty and diversity. Through correlation analysis based on multi-omics data, we observe a high prevalence of antiphage defense-related and phage-related protein families that co-occurred or co-expressed with RiPP families. Based on this putative association between RiPPs and phage infection, we constructed an Ocean Virus Database (OVD) and established a RiPP-involving host-phage interaction network through host prediction and co-expression analysis, revealing complex connectivities linking RiPP-encoding prokaryotes, RiPP families, viral protein families, and phages. These findings highlight the potential of RiPP families involved in prokaryote-phage interactions and coevolution, providing insights into their ecological functions in the ocean microbiome. Conclusions This study provides a systematic investigation of the biosynthetic potential of RiPPs from the ocean microbiome at a global scale, shedding light on the essential insights into the ecological functions of RiPPs in prokaryote-phage interactions through the integration of deep learning approaches, multi-omics data, and host-phage connectivity. This study serves as a valuable example of exploring the ecological functions of bacterial secondary metabolites, particularly their associations with unexplored microbial interactions.