The prokaryotic virus community is represented on the International Committee on Taxonomy of Viruses (ICTV) by the Bacterial and Archaeal Viruses Subcommittee. In 2008, the three caudoviral families Myoviridae, Podoviridae, and Siphoviridae included only 18 genera and 36 species. Under the able chairmanship of Rob Lavigne (KU Leuven, Belgium), major advances were made in the classification of prokaryotic viruses and the order Caudovirales was expanded dramatically, to reflect the genome-based relationships between phages. Today, the order includes six subfamilies, 80 genera, and 441 species. This year, additional changes in prokaryotic virus taxonomy have been brought forward under the new subcommittee chair, Andrew M. Kropinski (University of Guelph, Canada). These changes are:

  1. 1.

    replacement of “ phage ” with “ virus ” in prokaryotic virus taxon names. In recognition of the fact that phages are first and foremost genuine viruses, and to adhere to ICTV’s International Code of Virus Classification and Nomenclature (ICVCN), the word “phage” will disappear from taxon names, but not from phage names. For instance, the current taxon Escherichia phage T4 will be renamed Escherichia virus T4, while the name of this taxon’s member will remain unchanged (Escherichia phage T4). It is important that the community remembers the ICVCN distinction between viral taxa (such as species, genera, families, or orders) and their members, the actual viruses/phages: “viruses are real physical entities produced by biological evolution and genetics, whereas virus species and higher taxa are abstract concepts produced by rational thought and logic”;

  2. 2.

    elimination of the infix “ like ” from prokaryotic virus genus names. The naming of phage taxa has been an evolving process with genus names in the form “P22-like virus”, which was always considered to be a stop-gap measure, being replaced by P22likevirus. However, the latter convention is also problematic since it was only applied to genera included in the order Caudovirales, and the infix “like” was unnecessary since the grouping of viruses in genera implies per se that their constituent members are alike. Consequently, the infix “like” will be removed from the names of phage genera and genus names such as Lambdalikevirus and T4likevirus will become Lambdavirus and T4virus, respectively. It will of course remain correct to refer to “lambda-like viruses” and “T4-like phages” during discussions regarding specific groups of phages classified in these taxa. There have also been discussions in the Subcommittee whether all prokaryotic virus genera should adopt the system used for some archaeal and eukaryotic viruses, in which names of genera are created from the root of the corresponding family name with sequentially appended transliterated Greek letters (e.g., Alphabaculovirus, Betabaculovirus, etc.). However, it was decided that recognition of new genus names is of paramount importance and that further drastic changes in one setting might overly confuse the community. Thus, in most cases, the infix “like” was merely removed and the name of the founding member of the genus was retained as a root of the taxon name;

  3. 3.

    discontinuation of the use of “ Phi ” and other transliterated Greek letters in the naming of new prokaryotic virus genera. Since some scientists are under the impression that “Phi” in its various forms (phi, φ, Φ) indicates a phage, over the years, many phages were given names containing the prefix “Phi”. However, the prefix “Phi” adds no informational value when naming phage genera. Consequently, the Subcommittee decided that, unless there was sufficient historical precedent (e.g., Φ29 or ΦX174), Phi would no longer be added to genus names. In addition, Greek letters can create problems in electronic databases, as exemplified by a PubMed search for references on Bacillus phage Φ29 [1], which retrieved articles on phi 29, phi29, Phi 29, Phi29, 29 phi, {phi}29, φ29, and φ29 phages. Therefore, the Subcommittee strongly discourages phage scientists from using Phi or any other Greek letter in virus and virus taxon names in the future;

  4. 4.

    elimination of hyphens from taxon names. The ICVCN discourages hyphens in virus taxon names. Accordingly, taxon names such as Yersinia phage L-413C have been renamed (in this instance to Yersinia virus L413C). However, hyphens are retained when appearing in a number string: Thermus phage P2345 becomes Thermus virus P23-45 (its correct name) [2].

  5. 5.

    inclusion of the isolation host name in the taxon name. On several occasions, terms such as “Enterobacteria” or “Pseudomonad” have been used in phage taxon names. However, such terms do not refer to a specific bacterial host; nor do they indicate whether the phage in question was tested upon a variety of members of the particular host group. To improve the situation, terms such as “Enterobacteria” or “Pseudomonad” in taxon names will be replaced with the isolation host genus name: for instance, Enterobacteria phage T7 will become Escherichia virus T7. In addition, host species names will be eliminated from taxon names. For example, Thermus thermophilus phage IN93 will become Thermus virus IN93.

Further considerations

DNA-DNA relatedness is the gold standard in the classification of all prokaryotes [3–7], and efforts are underway to move towards a completely genomic taxonomy in that field [8]. The Bacterial and Archaeal Viruses Subcommittee has previously used overall proteome similarity to define genera and subfamilies, with 40 % homologous proteins indicating membership in the same genus [9–11]. This has resulted in spurious taxonomic lumping [12–14]. Furthermore, EMBOSS Stretcher [15, 16], which has been used for calculating nucleotide similarities between related phages (e.g., [17]), suffers from certain limitations (in particular the requirement for the genomes to be collinear). Problems with EMBOSS Stretcher are highlighted when an alignment of the phage T7 genome with a randomly shuffled T7 DNA sequence (http://www.bioinformatics.org/sms2/shuffle_dna.html) is attempted. The resulting value, 47.6 % identity, demonstrates that EMBOSS Stretcher values below a certain threshold are meaningless. Accordingly, more recent phage classification efforts have explored alternative approaches. Specifically, BLASTN [19] was found to be superior to EMBOSS Stretcher for identification and quantitative comparison of closely related phages [16]. Indeed, a BLASTN search seeded with the shuffled sequence of phage T7 specifically against “Enterobacteria phage T7” results in no detectable similarity, as expected from a randomized sequence with 48.4 % GC content. Moreover, BLASTN has also been used to determine relationships between phages at larger phylogenetic distances [17, 18], although the meaning of a similarity search hit in the absence of a true-shared ancestry remains unclear. Most of the newer programs that calculate phylogenetic relationships between genome sequences, including CLANS [20], GEGENEES [21], and mVISTA [22], are based upon sequence similarity analyses such as provided by BLASTN [19]. Complete and near-complete viral genome and protein homologies will be the focus of the Bacterial and Archaeal Viruses Subcommittee’s attention in 2016 to develop clearer parameters for the molecular definition of genera, subfamilies, and families.

The changes described here were formalized and submitted in more than 40 ICTV taxonomic proposals (TaxoProps) for consideration by the ICTV Executive Committee (http://www.ictvonline.org/). One new archaeal virus family (Pleolipoviridae), four new bacterial subfamilies (Guernseyvirinae [Salmonella phage Jersey], Vequintavirinae [Escherichia phage rV5], Tunavirinae [Escherichia phage T1], and Bullavirinae [Escherichia phage ΦX174]), and 59 new genera including 232 species are covered in these proposals (summarized in Table 1).

Table 1 Taxonomy proposals describing new taxa (genera, subfamilies, families) submitted to the ICTV in 2015
Full size table

While the Bacterial and Archaeal Viruses Subcommittee is delighted with the progress described here, some 400–600 new genomes of novel phages are deposited to GenBank annually. Many of these may have to be assigned to novel species or higher taxa via the ICTV TaxoProp process. Phage classification will therefore remain a highly demanding and daunting process, unless a genomic taxonomy for viruses is embraced (see [8]). Although a taxonomy that is based on the genome sequence alone might be incorrect due to rampant genomic rearrangements in viruses [23], such an approach may turn out to be the only scalable solution.