Daniel Straub

Thanks for commenting here. I avoided this issue for another recent metagenome assembly with version 3.2.1 by using `--megahit_options "--min-contig-len 500"`. I am not sure whether thats really an edge...

Forgot with my most recent assembly to adjust minimal contig length settings and arrived again at that issue. The assembly has 87,633,230 contigs with >=200bp. That is quite annoying indeed....

Generally I would like to assemble as much as possible and drop as least as possible data. I agree that bins are most interesting, but the assembly itself can also...

> This is why I would only suggest filtering internally for read mapping -> binning - we should not emit filtered raw assemblies. Ah sorry I misunderstood. Yes I agree....

So how is it going here?

Thanks! > Maybe it would be good to have an adaption from tab to csv as well since with nf-core schema it already supports them. `--input` should already accept tsv,...

On the first glance there seem two problems. (1) nf-test output files are included but shouldnt. (2) Something with pulling upstream seems to have gone wrong, because [.nf-core.yml](https://github.com/nf-core/ampliseq/pull/826/files#diff-07d8e637f9a7a8dd61dc0e46317a39485a55a2c02f790c726178da8ab5b90664) attempts to...

I tried DivNet some time ago and couldnt get it to run, see [nf-core slack](https://nfcore.slack.com/archives/CEA7TBJGJ/p1657721777903479?thread_ts=1657542647.094949&cid=CEA7TBJGJ). Maybe it was updated though.

All taxonomic classification subworkflows (see [here](https://github.com/nf-core/ampliseq/blob/3b252d263d101879c7077eae94a7a3d714b051aa/workflows/ampliseq.nf#L404-L480)) use ASV sequence tables (fasta) after the various filtering steps (see e.g. [DADA2](https://github.com/nf-core/ampliseq/blob/3b252d263d101879c7077eae94a7a3d714b051aa/workflows/ampliseq.nf#L420)). In any case, a subworkflow such as [here](https://github.com/nf-core/ampliseq/blob/master/subworkflows/local/dada2_taxonomy_wf.nf) would be great....