Comprehensive analysis of phage genomes from diverse environments reveals their diversity, potential applications,...

Phages are ubiquitous and diverse, playing a key role in maintaining microbial ecosystem balance. However, their diversity, potential applications, and their interactions with hosts and other phages remain largely unexplored. To address this, we collected 59,652,008 putative viral genomes from our laboratory, 45 public viral datasets, and an integrated public viral genome database (IGN), covering seven habitats. We obtained 741,692 phage genomes with completeness ≥50% (PGD50), and most (93.83%, 695,938/741,692) of these phage genomes were classified into the Caudoviricetes class. We found that 158,522 species-level viral clusters that contained 28.96% (214,814/741,692) phage genomes without any known phage genomes in the IGN, indicating substantial novelty. Global phylogenetic trees for five iterations based on complete phage genomes significantly expanded the known diversity of the virosphere. Genome analysis revealed phage potential divergence with habitat types and highlighted the utilization of alternative genetic codes. Furthermore, 3D structural similarity searches demonstrated significant potential for annotating previously uncharacterized viral proteins. Analysis of CRISPR spacer inferred potential hosts of phages and competitive networks among phages, highlighting virulent phages as promising candidates for phage therapy against pathogenic bacteria. Intriguingly, diverse CRISPR-Cas systems were detected within phage genomes themselves, suggesting their enormous potential as novel gene editing tools. Collectively, this study provides a comprehensive phage genome resource, foundational for future research into phage-host and phage-phage interactions, phage therapy development, and the mining of next-generation genetic tools.