<!channel>: I've added a couple of filter steps to the pipeline (even if I had promised myself not to touch this until we had that discussion):

--min_nongap

(0.3 by default) removes gappy, pre-aligned sequences,

--min_profile_cover

(0.8) removes non-pre-aligned sequences, aligned with HMMER, that cover a low proportion of positions in the HMM profile and

--raxtax_filter_rank

(1 == species by default) that removes sequences after running raxtax. I attach the output of the pipeline after running the latter two on GTDB's archaeal 16S sequences (with raxtax at genus, i.e. 2). At the moment, I don't have the resources to run the actual Sativa part that is downstream of the filters. For the same reason, I couldn't run the bacterial sequences at all since the alignment step takes too long. Will do when I can. All steps are now skippable, but on by default. See the pretty metro-map Claude made for me in the README.md: github.com/erikrikarddaniel/nf-core-sativa/tree/dev. (I have not proofread the docs yet.)