Overview

Teaching: min
Exercises: 240 min

Objectives

• Apply gene calling and functional annotation tools (Pharokka, Phold, Phynteny) to a real viral contig from your own assembly via the Phage Annotation Server

• Compare how much each tool contributes to annotating a genome

• Recognize why a large fraction of viral genes remain unannotated, even with modern tools

https://phage-annotation.org/jobs/34f6a6bdcbe64feea87639664a229ede/results

Choose a pet contig

In the previous lesson you learned about gene calling, and how Pharokka, Phold, and Phynteny use different approaches to annotate viral genes. Now you’ll apply all three to a contig from your own assembly.

Based on yesterday’s CheckV results, select a viral contig from your assembly that is either estimated High-quality or Complete. The TAs will help you coordinate so that everyone will work on a different contig.

The filtered viral sequences can be found in Viromics2026_workspace/1.3_virus_identification/30_filter_contigs/assembly.fasta

Exercise — extract your contig

• Use 1.4_10_select_contig.sbatch to extract a single contig from your filtered assembly. Fill in your assigned contig ID and submit it.

The Phage Annotation Server is a website, so you’ll need the fasta file on your own laptop, not just on Draco.

• In the VS Code Explorer, right-click on your extracted fasta file and select Download. Put somewhere you can find it on your laptop (e.g. your Downloads folder).

• Alternatively, you can open the fasta file containing the filtered viral sequences you created yesterday, and copy and paste the sequence to a text file on your own computer.

• If you found 20 or less viral sequences in your sample in the previous day, you can submit all viruses tpo phage-annotation.org (i.e. the complete Viromics2026_workspace/1.3_virus_identification/30_filter_contigs/assembly.fasta file.

You can check how many contigs are in this file with:

cat Viromics2026_workspace/1.3_virus_identification/30_filter_contigs/assembly.fasta | grep ">" | wc -l

Annotate your contig

Exercise — submit for annotation

• Go to www.phage-annotation.org and upload your virus(es).
• Leave every option on its default setting and submit.
• Supply an email address so that you’ll receive an update. The typical turnaround is a few minutes, depending on queue load.

Predict before you look

While you wait: roughly what fraction of your contig’s genes do you expect Pharokka alone (sequence-based) to confidently annotate? Do you expect Phold and Phynteny to meaningfully increase that number? Why or why not, given what each tool actually uses as evidence?

Now look at the output from the phage-annotation website. You might want to download the results, and use Excel, R, or Python to answer the following questions:

Exercise — interpret your annotation

1. How many genes were called on your contig in total?
2. How many were annotated by Pharokka’s sequence-based search alone? How many additional genes did Phold and/or Phynteny resolve? Does this match your prediction above? Look at the figure on the Phynteny website. Does this match your results?
3. Identify at least 2 structural genes (e.g. capsid, tail, portal). Where are they in the genome, and which tool provided the evidence for each?
4. Identify one replication-associated gene. What’s its function, and where in the phage life cycle is it used?
5. Note any other interesting features (CRISPR arrays, tRNAs, anti-CRISPR genes, auxiliary metabolic genes).
6. Based on everything above, what can you infer about this virus’s lifestyle (lytic vs. temperate) and its likely interaction with a host?
7. Look at the genome organisation. Describe what you see. Look for genes annotated by Phynteny but not any of the other tool. Do you notice anything particular?

Interpret your results

You’ll find that most genes on viral contigs are not actually annotated. There are several reasons:

1. Rapid evolution: Viruses can evolve rapidly, making homology-based annotation genuinely difficult. A gene may diverge beyond the point where detectable sequence similarity is possible.
2. Highly mosaic genomes: Phages frequently acquire and lose genes from their hosts and from other phages. As a result, related viruses can have very different gene complements, and reference databases do not fully capture this diversity.
3. Limited characterisation of viral sequence space: A large fraction of viral diversity remains unexplored. A hypothetical protein annotation is therefore not necessarily a failure of the annotation tools, instead it reflects how little of viral sequence space has been characterised.

Finally, download your phage-annotation.org results from the website. You will need these tomorrow

(Press the ‘Download all (.zip file) button in the top right corner.)

Key Points

• Most genes on viral contigs remain unannotated due to rapid viral evolution, highly mosaic genomes, and how little of viral sequence space is characterized in reference databases

• Using structure-based search or genomic context can be used to detect distant homologies, allowing for the annotation of more viral genes.