Background: With advances in sequencing technology and decreasing costs, the number of phage genomes that have been sequenced has increased markedly in the past decade. Materials and Methods: We developed an automated retrieval and analysis system for phage genomes (https:// github.com/RyanCook94/inphared) to produce the INfrastructure for a PHAge REference Database (INPHARED) of phage genomes and associated metadata. Results: As of January 2021, 14,244 complete phage genomes have been sequenced. The INPHARED data set is dominated by phages that infect a small number of bacterial genera, with 75% of phages isolated on only 30 bacterial genera. There is further bias, with significantly more lytic phage genomes (*70%) than temperate (*30%) within our database. Collectively, this results in *54% of temperate phage genomes originating from just three host genera. With much debate on the carriage of antibiotic resistance genes and their potential safety in phage therapy, we searched for putative antibiotic resistance genes. Frequency of antibiotic resistance gene carriage was found to be higher in temperate phages than in lytic phages and again varied with host. Conclusions: Given the bias of currently sequenced phage genomes, we suggest to fully understand phage diversity, efforts should be made to isolate and sequence a larger number of phages, in particular temperate phages, from a greater diversity of hosts.

Clostridium difficile infection (CDI) is a major cause of infectious diarrhea. Conventional antibiotics are not universally effective for all ribotypes, and can trigger dysbiosis, resistance and recurrent infection. Thus, novel therapeutics are needed to replace and/or supplement the current antibiotics. Here, we describe the activity of an optimised 4-phage cocktail to clear cultures of a clinical ribotype 014/020 strain in fermentation vessels spiked with combined fecal slurries from four healthy volunteers. After 5 h, we observed ~6-log reductions in C. difficile abundance in the prophylaxis regimen and complete C. difficile eradication after 24 h following prophylactic or remedial regimens. Viability assays revealed that commensal enterococci, bifidobacteria, lactobacilli, total anaerobes, and enterobacteria were not affected by either regimens, but a ~2-log increase in the enterobacteria, lactobacilli, and total anaerobe abundance was seen in the phage-only-treated vessel compared to other treatments. The impact of the phage treatments on components of the microbiota was further assayed using metagenomic analysis. Together, our data supports the therapeutic application of our optimised phage cocktail to treat CDI. Also, the increase in specific commensals observed in the phage-treated control could prevent further colonisation of C. difficile, and thus provide protection from infection being able to establish.