Integration of pyrodigal-gv for viral gene prediction in Metabuli

[Enkabloza]

Hi Kimbeom,

I hope you are doing well. I have a question and a suggestion regarding the possibility of expanding Metabuli’s functionality for viral research.

Would it be possible to add support or an optional module that allows Metabuli to use pyrodigal-gv (https://github.com/althonos/pyrodigal-gv) instead of the standard Prodigal?
Since Metabuli currently relies on Prodigal for gene prediction, integrating pyrodigal-gv (which includes specialized models for giant viruses and viruses using alternative genetic codes) could potentially improve the detection of viral genes, especially in metagenomic datasets enriched with NCLDVs or phages using genetic code 15.

For reference, pyrodigal-gv is a Python-based extension of Prodigal that includes models for:

Acanthamoeba polyphaga mimivirus

Paramecium bursaria Chlorella virus

Acanthocystis turfacea Chlorella virus

VirSorter2 NCLDV model

Gut phage models (genetic codes 11 and 15)

It could be very useful if Metabuli could either accept amino acid sequences as input or internally switch to pyrodigal-gv for viral genome annotation.

Thank you very much for your time and for all your great work on Metabuli. I’d really appreciate your thoughts on whether this could be feasible in future versions.

Best regards

[jaebeom-kim]

Hi, Thank you for this suggestion!
You can provide cds information to skip Prodigal's gene predictions.
It can be done using --cds-info option to provide Genbank/RefSeq-format CDS files.

I will put using pyrodigal-gv for viruses in my todo list.