archeal genomes about bakta

[richardstoeckl]

Actually, Bakta was designed and developed to annotate isolated bacterial genomes. In particular, the database comprises bacterial protein coding genes, only. Also, there are currently no options/parameters to surpass archaeal taxonomic information to 3rd party tools of the workflow, e.g. tRNAScan-SE.

Since bakta now has several huge advantages compared to prokka (e.g. pyrodigal and circos plots), is there any chance for an archaea-mode in the future, or is this design decision final?

[oschwengers]

Actually, Bakta was designed and developed to annotate isolated bacterial genomes. In particular, the database comprises bacterial protein coding genes, only. Also, there are currently no options/parameters to surpass archaeal taxonomic information to 3rd party tools of the workflow, e.g. tRNAScan-SE.

Regarding metagenomes:
For sure, a very interesting question. I haven't tried it yet but I'll put it on the list and will look into this.

[oschwengers]

Hey @richardstoeckl, thanks for this follow-up question which gets asked more and more often. I totally see the point in this - I just haven't come up with a decision whether it's better to either add an archaea mode to Bakta or to start a new distinct project, maybe archa?

A distinct project would be highly-inspired by Bakta itself sharing a lot of code but would provide room for important distinctions we might have to make, both now and in the future. However, this would induce a lot of code duplication which could only be mitigated and handled by starting a 3rd library project.

In conclusion, yes there is definitely a chance for a "bakta-like" archaea annotation solution - I'm just not yet quite sure what might be the best way how to do it. For now, I rather tend to the 2nd idea of a distinct project.

[oschwengers]

A quick follow-up:
Yes, Bakta works absolutely fine on bacterial metagenome-assembled genomes. Actually, due to its large and taxonomically untargeted database, it has a competitive edge compared to Prokka.