Page 22 - Read Online

Page 16 of 21 O’Connell et al. Microbiome Res Rep 2023;2:21 https://dx.doi.org/10.20517/mrr.2023.17

of these genera mirror the monophyletic nature of the genera Unicornvirus and Amginevirus, which were
included for comparative purposes.

Identifying monophyletic groups in the VICTOR analyses also proved extremely important for untangling
subclusters K1 and K6 and redefining the boundaries of these groups. When the VIRIDIC output is
compared to the subcluster assignments, it is clearly illustrated that K1 and K6 are intertwined with each
other, while the remaining subclusters are quite distinct from each other [Figure 8A]. A similar (though not
identical) intermingling is seen when the previous Gegenees alignment is compared with the subcluster
assignments [Figure 8B]. As established for previous clusters, it appears that approximately 50% proteome
similarity is quite a robust metric for the identification of subclusters. Therefore, the creation of several
novel subcluster assignments would likely better reflect the genetic diversity of these MP highlighted in the
VIRIDIC and Gegenees outputs. Based on the VICTOR analysis and the observation that subclusters form
monophyletic groups, it appears that there is sufficient support to warrant the addition of nine novel
subclusters in total [Figure 8C]. These additional subclusters would support the hypothesis that one genus
can be assigned to a single subcluster, which in this case helps reflect the diversity of the genera that only
feature one MP within this dataset.

DISCUSSION
Phages will most likely prove an essential part of the effort to overcome the very concerning threat of
antibiotic resistance. As the most abundant biological entities on Earth , an overwhelming arsenal is
[29]
hypothetically available for the design of phage-based therapeutics and diagnostics, and the possibility to
genetically engineer the phages makes the composition of phage cocktails endless. In order to capitalise
upon the diversity of phages, it is important to have robust classification systems in place. Lawrence et al.
described in immense and commendable detail the difficulties of applying traditional Linnaean
classification (essentially traditional hierarchal taxonomy based on shared characteristics) with particular
regard for how this style of classification underrepresents the diversity of phages, especially when genetic
mosaicism is considered . Many of the concerns raised by Lawrence et al. have been satisfied by the
[7]
[7]
[12]
roadmap for genome-based taxonomy proposed by Turner et al. . The roadmap recommends the
abolishment of many of the Linnaean-type classifications and recommendation that whole genomes (as
[10]
opposed to a core genome) be considered when assigning groups. VIRIDIC (which predicts genus and
species assignments based on pairwise nucleotide comparisons with consideration for genome length and
aligned genome fraction) has been heavily employed in a massive undertaking to update the taxonomy of
phages according to the Turner et al. (2021) proposal (ICTV Master Species List; https://talk.ictvonline.org/
files/master-species-lists/m/msl/12314). For this reason, VIRIDIC was selected as the initial analysis to
identify groups of MP that may belong to novel genera.

After the identification of potentially novel genera in this study by comparing the VIRIDIC results to the
existing taxonomic information, the genomes belonging to the predicted genera were analysed with
Gegenees . The reasoning for performing this analysis is that (hypothetically) the protein-based
[21]
alignments would reflect the VIRIDIC (i.e., nucleotide) alignments and provide additional support and
further confirm the predicted genera (although it should be noted that the genetic mosaicism, temperate
lifestyle and limited host range of MP often makes nucleotide and proteomic analyses complicated; ). If
[24]
supported by both the VIRIDIC and Gegenees outputs, phylogenetic trees were expected to illustrate the
novel genera as monophyletic branches . Following this workflow, 20 well-supported novel genera were
[12]
identified across three clusters, A, J and K. In one instance, evidence supporting the abolition of a genus in
cluster G (Antsirabevirus) was presented. These genera were mostly identified following the widely accepted
≥ 70% nucleotide similarity demarcation proposed by Turner et al. which was heavily employed in the