Phage Annotation Server

Annotate your phage genome

Upload a nucleotide FASTA file and we'll run pharokka → phold → phynteny for you: gene calling, structure-guided functional annotation, and synteny-based completion, all in one job. Free, for research use.

Every job runs all three, in order:

Pharokka logo Pharokka Phold logo Phold Phynteny logo Phynteny

Nucleotide FASTA, single or multi-contig (up to 20 contigs, 2,000,000 bp total). Up to 50 MB — .gz / .bz2 / .zst compressed files are fine.

Only used to send you your results link. No account, no mailing list.

Advanced options (pharokka / phold / phynteny parameters)

Every real CLI flag is shown next to its label, with the tool's own default. Leave a field as-is to use that default. Values are re-validated on the server regardless of what's shown here.

Metagenome mode

Auto turns metagenome mode on when your upload has 2 or more contigs (as recommended by pharokka), and off for a single contig. pharokka itself refuses --meta with only 1 contig, so "Always on" is automatically skipped in that case -- we'll tell you what was actually used on the job status page.

Pharokka

Default 1e-05. Allowed range 1e-30–0.1.

Default 8.5. Allowed range 1.0–10.0.

Phold

Default 0.001. Allowed range 1e-30–0.1.

Default 9.5. Allowed range 1.0–10.0.

Phynteny

Default 0.8 (0.0–1.0). Predictions below this confidence are reported but not treated as high-confidence calls.

Don't have a genome handy?

Runs the complete pipeline on NC_043029, a small complete phage genome (7.6 kb, 17 CDS). Takes about a minute and produces a real result you can explore — not a canned demo.

Prefer to script it? Everything here is available through the API, with a ready-to-run Python client.

By submitting you agree to the terms of use and accept that uploaded sequence data will be processed on shared infrastructure.

What happens to your genome

  1. Pharokka

    Rapid gene calling and functional annotation against phage-specific databases (PHROGs, CARD, VFDB, and more).

  2. Phold

    Structure-guided annotation with a protein language model, recovering function for genes pharokka couldn't annotate.

  3. Phynteny

    Fills in remaining unknown genes using synteny patterns learned across thousands of phage genomes.

Typical turnaround is a few minutes to around twenty minutes per genome, depending on queue load. You'll get a shareable status link the moment you submit -- and it's also saved to My jobs in this browser.

704 phage genomes annotated so far.

Supported by

The compute and storage behind this service are provided by the ARDC Nectar Research Cloud, a national research infrastructure service supported by the Australian Government through the National Collaborative Research Infrastructure Strategy (NCRIS). This service would not be free to use without them.