Unraveling the diversity of prokaryotic immune systems by in silico comparative analysis

Bacteria and Archaea have evolved a variety of antiphage immune systems including the well-known Restriction-Modification (RM) and CRISPR-Cas, and several recently discovered systems. Nevertheless, which traits are shared between these systems and how their diversity is originated are underexplored questions. Thus, we sought to compare the immune system family diversity using a robust sequence similarity approach and a network to analysis. The immune systems were predicted using Defense-Finder followed by a robust sequence comparison of all-vs-all immune system sequences (instances) using HH-suite. A network was built, the properties of the network were determined, and the protein domains associated to the systems were identified. We obtained a network of 15,289 nodes (instances) distributed among 82 immune system families. The network highlighted several conserved traits shared between instances belonging to different families and one big cluster containing ~90% of the instances of 45 immune system families. These families comprise the most known family RM as well as the less widely distributed CBASS, BREX, Lamassu, Gabija, and Wadjet families. The network centrality analysis showed that the median of the instances belongs to RM type I, Hachiman and PsyrTA families being the most central in the network. In contrast, several of the newly reported families such as Thoeris, Druantia and Zorya among others are distributed in the network peripheral region. We found that more than 80% of the relationship revealed in the network reflect at least one known shared protein domain related to DNA metabolism such as helicases and nucleases. Our results show that behind the prokaryotic immune system diversity, there are several conserved traits. We also show the importance of RM type I to support the network of 45 immune systems, suggesting this type of RM could be key in the generation of the observed diversity.