Page 15 - Read Online

O’Connell et al. Microbiome Res Rep 2023;2:21 https://dx.doi.org/10.20517/mrr.2023.17 Page 9 of 21

resolution of genera could be obtained based on a ≥ 50% proteome similarity threshold, which reduces the
number of novel genera to ten [Figure 2B]. The ten genera were further supported by the VICTOR output,
as each demonstrated a monophyletic nature similar to the genus Luchadorvirus which was included for
comparative purposes [Figure 2C]. The suggested nomenclature for the novel groups is indicated in
Figure 2C. Seven of these genera provide additional support to the proposed genus-subcluster relationship,
with subclusters A8 and A10-13 each assigned to a single genus. The exceptions were subclusters A9 and
A2. Of the five A9 MP included in the analysis, phage Yecey3 was assigned to a different genus
(Yecey3virus) compared to the remaining four (Almavirus). Meanwhile, the eight A2 viruses were grouped
into three genera (Adzzyvirus, D29virus, and Serenityvirus; Figure 2C).

Interestingly, within the VIRIDIC alignment, subcluster A2 is “interrupted” by the A17 genome
[Figure 3A]. This is unexpected as subcluster groups are based on “recognisable divisions” within nucleotide
similarity alignments, so it would be anticipated that each subcluster would align together in a defined
group. Therefore, this interruption would suggest that there is a discrepancy in the organisation of A2.
Considering a nucleotide similarity threshold of ≥ 60%, there are three groups apparent within the VIRIDIC
alignment of the A2 phages, which may be better represented by three subclusters [Figure 3A]. These
groups are also reflected in the Gegenees proteome alignment of these phages when a proteome similarity
threshold of ≥ 50% is applied [Figure 3B]. The VICTOR-generated phylogeny did in fact reveal that the
current A2 subcluster is comprised of three monophyletic branches which support the groups identified in
the previous alignments [Figure 3Ci]. This is quite unusual, as the other A subclusters included in the
phylogeny have demonstrated that subclusters are monophyletic. Therefore, it is reasonable to propose the
creation of two novel subclusters from two of the branches, with the largest branch remaining as A2, in
order to implement as little reorganisation as possible [Figure 3Cii]. This would create a single subcluster
for Pukovnikviruses and Turbidoviruses (A2), and the second subcluster would encompass the novel genera
Adzzyvirus, D29virus, and Serenityvirus.

Cluster G - Supporting evidence for the removal of subcluster G5
Cluster G is a relatively small cluster featuring five subclusters and 18 genomes within this dataset. VIRIDIC
did not identify any novel genera within this group, but rather assigned four genera to the five subclusters.
Three of these genera reflected the G1, G2 and G4 subclusters (thereby supporting a relationship between
genus and subcluster), whereas G3 and G5 were assigned a single genus based on the established ≥ 70%
similarity demarcation [Figure 4A]. This contradicts the existing taxonomy of these latter subclusters, as the
G5 MP, Antsirabe, belongs to its own genus, Antsirabevirus [Supplementary Table 1]. However, the
Gegenees output for this cluster indicates that even at a protein level, a single genus for the G3 and G5 MP
is supported based on the proposed ≥ 50% proteome similarity threshold [Figure 4B]. This threshold also
supports the removal of Antsirabe from G5 and its inclusion within G3, as this similarity boundary has
supported the novel subclusters proposed in previous sections [Figure 3]. Further support for assigning a
single genus and a single subcluster to these two subclusters is provided by the VICTOR-generated
phylogenetic tree, as the G3 and G5 are shown to share a singular branch, similar to the mono-phylogeny of
the other genera and subclusters [Figure 4C].

Cluster H - One novel subcluster
Within this dataset, cluster H is comprised of six phages that are organised into three genera and two
subclusters. Subcluster H1 features two genera, Predatorvirus and Konstantinevirus. Predatorvirus is
represented by a single MP, Predator. This phage does not meet the ≥ 70% nucleotide similarity required for
inclusion with Konstantinevirus [Figure 5A]. The Gegenees analysis of this cluster also highlights that
Predator does not meet the proposed ≥ 50% proteome similarity for inclusion with Konstantinevirus, which