I'm very excited to announce that in the next few weeks, we'll be launching the new Actinobacter phage database which will merge the existing Myco, Arthro, and Strepto DBs as well as have the ability to hold information about any phage that infects a host within the Actinobacter Phylum. Woo!
PhagesDB started as a way to organize and share information about the growing number of mycobacteriophages coming out of the University of Pittsburgh's Hatfull lab and PHIRE program. Spreadsheets weren't cutting it any more, and so a database-driven website seemed like the way to go. With the introduction of HHMI's SEA-PHAGES program, the number of Mycobacteriophages being isolated and sequenced has grown exponentially the past several years, and PhagesDB has been indispensable in collecting and sharing this information.
Over the past couple years, both the Hatfull lab and the larger SEA-PHAGES program have sought out new host bacteria to use in the isolation of phages. The first "official" host other than M. smegmatis mc2155 to be rolled out to the SEA-PHAGES program was an Arthrobacter sp. strain. When SEA-PHAGES schools began using this strain, it became imperative to have a place like the Mycobacteriophage database to record and share their data. Thus, the Athrobacter phage database was born, as was, for similar reasons, the Streptomyces phage database.
One DB or many DBs?
As the new hosts hit the bench, we discussed whether to keep phages in the same database or split them by genus. We decided on the latter, because Clusters were genus-specific, because the existing software architecture was designed that way, and because the Mycobacteriophage_Draft Phamerator database was already taking two weeks to update; the idea of trying to stuff more phages into the same box seemed impossible.
Now we're reversing course, and soon the Mycobacteriophage, Streptomyces phage, and Arthrobacter phage databases will be replaced by a single database holding phages that infect bacterial hosts within the Phylum Actinobacter. This is possible due to the new kClust method of Phameration, the decision to make Clusters span multiple genera, and the fact that I have a little time to write the code. It's necessary because we and the SEA-PHAGES program plan on adopting dozens of new Actinobacter hosts in the coming years. Making and maintaining separate DBs for each genus is no longer feasible. (Or smart!)
What and When
At some point in the next few weeks, before the 7th Annual SEA-PHAGES Symposium, the phagesdb.org URLs will point to the ActinobacterDB instead of the Mycobacteriophage DB. Host information will be more prominently displayed across the site, since we can no longer assume that all phages infect the same genus!
We've also made the decision that the clusters we know and love will stay the same, but they will be known as Actinobacter phage clusters rather than Mycobacteriophage clusters. This means that while Cluster V contains phages that infect Mycobacterial hosts, Cluster W could contain phages that infect Gordonia, and Cluster X could hold phages that infect Streptomyces. This also allows phages that infect different genera to be contained in the same cluster...no need to artificially separate them by their isolation host. Eagle-eyed users of PhagesDB may have already noticed that the Cluster A phage KatherineG was isolated on a Gordonia strain, not a Mycobacterium strain.
We'll pass 1000 sequenced Actinobacter phages any day now, and are excited to have multiple genera and species reach the point we've reached with M. smegmatis. The recent elife paper describes many of the wonderful things you can do with a large dataset of phages that infect a single species. Imagine what we can discover together with a dataset that includes so many host genera, species, and strains. As always, this will be a community effort, and we need your help to continue discovering new phages, analyzing new genomes, and entering data. Dare I suggest...a future Actinobacter phage paper with 2,000 phages? 5,000? 10,000?! I can't wait to see.