Fig. 5

Characterization of the virus-encoded auxiliary metabolic gene (AMG) epsG. A Genome map of the CB virus vOTU4_CB17_43158 encoding epsG. This vOTU had a circular genome due to the overlapped regions between two ends. The full genome was recovered by long-read assembly, while the two contig fragments, which had 100% DNA identity to the full genome, were recovered by SR assembly. CheckV was used to assess host-virus boundaries and remove potential host fractions on the viral contig (no host contamination was found; no cellular gene, except the AMG, was identified; see “Methods”). Genes were marked by four colors to illustrate AMGs (red), phage structural genes (blue), other phage genes (orange), and unaffiliated genes (gray). AMGs were detected by DRAM-v and following manual inspection; phage genes were classified by comparing their predicted protein sequences to those of a large database of 15,958 profile hidden Markov models by CheckV and of viral genes in the extended RefSeqABVir database by VirSorter v1 in virome decontamination mode. Genes were also annotated by comparing them to UniRef, PFAM, and KEGG databases. Genes were marked as “phage genes” if they were matched to the genes of viruses in the RefSeqABVir database or CheckV databases. Genes were considered “unaffiliated genes” if they had no hit to a sequence in any tested databases. B Phylogenetic tree of the viral and microbial epsGs. The tree was inferred using maximum likelihood (ML) method with EpsG protein sequences (see “Methods”). Bootstrap values (expressed as percentages of 1000 replications) of ≥40 are shown at the branch points. The scale bar indicates a distance of 0.5. The vEpsG sequence is indicated in red. The mEpsG sequences from the CB microbial metagenomes [10] and the NCBI nr database are indicated in purple and black, respectively. The full phylogenetic tree without collapse is provided in Figure S6