PSOSP uncovers pervasive SOS-independent prophages with distinct genomic and host traits in bacterial genomes.

Prophages are ubiquitously present in bacterial genomes, significantly influencing host physiological and ecological functions. We present Prophage SOS-dependency Predictor (PSOSP), a novel bioinformatics tool that predicts prophages induction modes by analyzing the Heterology Index (HI) of LexA protein binding to target DNA, classifying prophages into SOS-dependent prophages (SdPs) and SOS-independent prophages (SiPs). PSOSP was experimentally validated to accurately distinguish SdPs from SiPs in a test set, achieving 100% sensitivity and specificity. Applying PSOSP to 49,333 complete bacterial genomes, we inferred prophage induction modes in 15,493 bacterial genomes and identified 11,806 SiPs. These SiPs are widely distributed across 145 bacterial genera and exhibit distinct genomic features compared to SdPs. Correspondingly, the hosts of these two prophage types are hypothesized to differ in their physiological characteristics. These findings, powered by PSOSP, provide not only novel insights into the diverse induction mechanisms but also a critical methodology for future phage-host interaction studies.