Before assembling anything, always look at your test genome. Make sure you’re in ~/p/input.

# activate the QC environment
conda activate qc_env
ls
seqfu stats *.fq

Save the stats to a file instead of just printing them:

seqfu stats *.fq > stats.tsv
ls
head stats.tsv
conda deactivate

Read more about SeqFu: telatin.github.io/seqfu2

Shovill wraps SPAdes (and other assemblers) with sensible defaults for small, high-coverage genomes like phages — much faster than running SPAdes directly. Still inside ~/phage_training/input:

conda activate shovill

shovill --outdir ./assembly \
  --R1 phage_R1.fq \
  --R2 phage_R2.fq \
  --cpus 4

ls
conda deactivate

Your assembled contigs will land in ~/phage_training/input/assembly/contigs.fa.

Pharokka is a rapid annotation tool built specifically for bacteriophage genomes — gene calling, functional annotation against phage-specific databases, tRNA/CRISPR detection, and more.

conda activate pharokka

pharokka.py -i Shovill_Contigs.fasta -o ./pharokka -t 4

ls
ls pharokka

Have a look at the outputs — in particular pharokka.gbk, pharokka.gff, and the summary file.

Pharokka ships with its own plotting utility, generating a circular genome map — genes coloured by functional category, strand, tRNAs, CRISPRs, and more. A great figure for a report or paper. Still inside ~/phage_training/input, with the pharokka environment active:

pharokka_plotter.py -i Shovill_Contigs.fasta -o ./pharokka \
  -n pharokka_plot --gff pharokka/pharokka.gff

ls pharokka

Copy the finished plot out to Windows so you can open it:

cp pharokka/pharokka_plot.* /mnt/c/Users/USER/Downloads/
conda deactivate

Phold builds on Pharokka’s output, using structure-based annotation to assign function to the many phage genes with no similarity to anything in sequence databases (“hypothetical proteins”). This is one of the most powerful things you can currently do for phage genomes, since so much of phage diversity has no close sequence homologs.

conda activate phold

phold run -i pharokka/pharokka.gbk -o phold -t 4

conda deactivate

Compare how many CDS went from “hypothetical protein” in Pharokka to having a predicted function in Phold — this is usually a striking difference.

Once you have more than one annotated phage genome (your own isolates, or reference genomes from NCBI/INPHARED), it’s useful to visualise how gene order/content is conserved or rearranged between them — this is “synteny.”

Phold’s built-in comparison mode works directly off its own outputs (only relevant once you have more than one sample’s phold folder to compare):

conda activate phold

phold compare -i phold_<sample_A> phold_<sample_B> \
  -o ./synteny_output \
  -t 4

conda deactivate

Alternatively, clinker compares gene clusters directly from GenBank files (e.g. your pharokka.gbk outputs):

conda activate clinker

clinker pharokka_<sample_A>/pharokka.gbk \
        pharokka_<sample_B>/pharokka.gbk \
  -p ./synteny_output/clinker.html

conda deactivate

VipTree builds a genome-wide proteomic tree from whole-genome amino acid content, rather than a single marker gene — well suited to phages, which often lack universal marker genes. No install needed; we use the web version.

🔗 www.genome.jp/viptree

Upload your assembled/annotated contig(s) (FASTA), select the appropriate reference dataset (e.g. “Prokaryotic dsDNA viruses”), and submit. This places your phage(s) in context against thousands of reference genomes based on overall proteome similarity.