Page 19 - Read Online

O’Connell et al. Microbiome Res Rep 2023;2:21 https://dx.doi.org/10.20517/mrr.2023.17 Page 13 of 21

would further suggest that the Predatorvirus is a well-established genus [Figure 5B]. The phylogeny of this
cluster features at least two very distinct branches that relate to subclusters H1 and H2 [Figure 5Ci].
Predator appears to be farther removed from the other H1 cohorts, and considering it fails to meet the
proposed proteome threshold, there is therefore an argument in favour of creating a new subcluster for
Predatorvirus [Figure 5Cii]. By creating this additional subcluster, the genus-subcluster hypothesis is
reinforced further, as the three genera within cluster H will be assigned to three subclusters, respectively.

Cluster J - One novel genera and two novel subclusters
Currently, cluster J is not subdivided into subclusters. This would suggest a lack of “recognisable divisions”
amongst the MP at a nucleotide level. However, VIRIDIC was able to allocate several genera to the 11
genomes included in the dataset from this cluster [Figure 6A], indicating there is at least a clear enough
division to make taxonomic assignments based on nucleotide similarity. The VIRIDIC output for cluster J
predicted 4 genera within this cluster based on the ≥ 70% similarity threshold [Figure 6A]. However,
following visual inspection, it was noted that applying a slightly lesser threshold of ≥ 66% allows for a more
noticeable distinction of two groups as opposed to four. The Gegenees output seems to agree with this, as
applying the proposed genus, a threshold of ≥ 50% proteome similarity appears to reflect these two groups
[Figure 6B]. As the VICTOR-generated dendrogram clearly illustrates two unique branches, it is arguable
that the most confidence can be placed in the existence of two genera within this cluster. As it is a goal of
this study to interfere with the existing taxonomic classifications as possible, it was decided to only
acknowledge these two genera despite the deviation from the ≥ 70% threshold. The proposed nomenclature
for these genera is Bakavirus and Omegavirus, as indicated in Figure 6C. Omegavirus is the established name
of the existing genus recognised in cluster J, so only Bakavirus may be considered a novel finding.

With regards to potentially creating subclusters for these genomes, the creation of two subclusters within
cluster J appears to be robustly supported, considering the VIRIDIC and Gegenees outputs highlight at least
two groups, the latter of which can be supported by the proposed subcluster threshold of ≥ 50% proteome
similarity. By creating a subcluster for each genus identified, i.e., J1 (Bakavirus) and J2 (Omegavirus), it
would better reflect the diversity of these phages, and it would further support the proposed notion that
genus assignment can be a predictor of subcluster formation and that ≥ 50% proteome similarity is a reliable
threshold for subcluster creation.

Cluster K - Eight novel genera and nine novel subclusters
In total, 66 cluster K phages were analysed to identify novel groups. Similarly to previous clusters, several K
subclusters support the hypothesis that one subcluster can be assigned to a single genus, but there were two
subclusters that disagreed with this observation. Based on the VIRIDIC analysis of this cluster, K1 and K6
appear to be comprised of several genera when considering the nucleotide similarity threshold of ≥ 70%
[Figure 7A]. MP Yunkel11 and Marshawn appear to be closely related to the “neighbouring” genus, and
visual inspection (similar to that applied for cluster J in the previous section) appears to support the
combination of a few of the VIRIDIC-predicted genera based on these highly related groups. However, the
subsequent Gegenees analysis of these phages [Figure 7B] provided clearer distinctions between these
groups when a proteome similarity threshold of approximately 60% is applied. This 60% threshold also
creates two groups out of the phages Ekdilam, Amohnition, DarthP, Amgine and Ellie, which VIRIDIC
(and the existing ICTV/NCBI taxonomy) classifies as a single genus. While reducing the threshold to the
previously adopted ≥ 50% recombines these two groups, it does not entirely clarify the boundaries between
Yunkel11 and Marshawn and their neighbouring genera. The VICTOR analysis, on the other hand, was able
to demonstrate the monophyletic nature of all the proposed genera, including that of Yunkel11 and
Marshawn, indicating 8 novel groups, the names of which are listed in Figure 7C. The monophyletic natures