Large-scale analysis of bacterial genomes reveals thousands of lytic phages

Phages are typically classified as temperate, integrating into host genomes, or lytic, replicating and killing bacteria; for this reason, lytic phages are not expected in bacterial genome sequences. Here we analyse 3.6 million bacterial genome assemblies from 1,226 species and find 119,510 lytic phage genomes, which we term bacterial assembly-associated phage sequences. This represents a ~5-fold increase in the number of phages with associated hosts and raises questions about fundamental aspects of phage biology. Our analyses of bacterial assembly-associated phage sequences revealed previously undescribed phage clusters, including clusters distantly related to Salmonella Goslarviruses in Escherichia coli and Shigella, while also substantially expanding known genera such as Seoulvirus (from 16 to >300 members). Close relatives of lytic phages used therapeutically were also detected, suggesting clinical isolate sequencing unknowingly archives potential phage candidates. The discovery of complete, lytic phage genomes within bacterial assemblies challenges assumptions about the nature of the lytic lifestyle and reveals an untapped reservoir of phages.