could the annotated sequence as the input for gapseq

[hongzhonglu]

Hi @jotech ,
Thanks for your nice work! I have a question about the input sequence format to build model using gapseq.
For example, I have 2000 annotated proteins in a genome, is this can be as input for gapseq to generate the metabolic model? Looking forward to your help!

Best regards,
Hongzhong

[Waschina]

Hi Hongzhong,

in the current Version, gapseq supports only the genome input as nucleotide fasta sequence. But I agree, a protein fasta for the annotated proteins as alternative input would be a useful extension. I will mark this as a feature request, which we will address in an upcoming version.
Until then, you could use the coding sequences of your proteins in a nucleotide fasta as gapseq input as a workaround.

Best regards,
Silvio

[hongzhonglu]

Thanks so much! Looking forward to the new functions.

[jdwinkler-lanzatech]

Yes, this would be great-BLASTP (or DIAMOND) would be much faster than TBLASTN based on my experience.

[Waschina]

Yes, we actually have plans in those directions; specifically to employ DIAMOND.
But since this is not per se an issue in the workflow, would it be okay with you to convert this conversion to the discussion forum (a new feature in github)?

[jdwinkler-lanzatech]

Definitely!

[silvtal]

Hi, thank you for developing gapseq

This feature has not been implemented, right? I have structural and functional annotations in various formats and I would love to be able to directly use them as input for ./gapseq draft. I'd simply re-annotate, but I launched ./gapseq find 11 and a half hours ago on my university cluster and it's still running.

If it hasn't been implemented, could you give me some guidelines on how to parse my data so I obtain the "minimum necessary" .tbl files for draft? I've seen you need both the reactions and the transporter files, which columns are absolutely necessary?

Thank you!