Documentation does not fit pipeline behaviour.

[oliverdrechsel]

Description of the bug

https://nf-co.re/differentialabundance/1.2.0/docs/usage mentions in section Contrasts file
blocking is listed under optional arguments.

Leaving out this column from contrasts file, the pipeline will complain

  Error in read_contrasts(contrasts_file, samples) :
    Contrasts file must contain all of variable, reference, target, blocking
  Calls: validate_inputs -> read_contrasts
  Execution halted

---

In addtion documentation's Quick start mentions
-profile rnaseq

using this profile will cause an error message that this profile is not defined

Unknown configuration profile: 'rnaseq'

Command used and terminal output

# this call fuses both issues

$ nextflow run nf-core/differentialabundance      \
--input samplesheet.csv      \
--contrasts contrasts_small.noblocking.csv      \
--matrix salmon.merged.gene_counts.tsv      \
--gtf references/Homo_sapiens.GRCh37.87.gtf      \
--outdir test_output      \
-profile rnaseq,rki_slurm,rki_singularity,singularity

Relevant files

# contrasts file

id,variable,reference,target,exclude_samples_col,exclude_samples_values
infected_time_comparison,incubation,15,24,condition,control

System information

N E X T F L O W ~ version 23.04.1
https://github.com/nf-core/differentialabundance [determined_waddington] DSL2 - revision: 3a849f0 [master]

[WackerO]

Hey there @oliverdrechsel

So regarding the unknown profile I am very sorry, but I cannot reproduce that error...I just ran the pipeline with that profile and it worked just fine; it is defined in the nextflow.config.

Regarding the problem with the contrasts file, your layout is not correct. I can see how the docu is a bit unclear in that regard, I will update it.
I'm not quite sure what the contrasts should look like in your case but I believe it might be this?

id,variable,reference,target,blocking,exclude_samples_col,exclude_samples_values
infected_time_comparison,incubation,15,24,,condition,control

That is to say, the header of the csv file has to be the same always; the entry for blocking can be left out, but the word blocking itself still has to appear in the header row. I'll specify this in the docs :)

I'm closing this for now, don't hesitate to reopen if you still have trouble with the pipeline!

[WackerO]

Just had a look at the code, my bad; the optional columns are treated a bit differently right now, i.e. exclude_samples_col and exclude_samples_values are truly optional whereas for blocking, at least the column name seems to be necessary right now. Will see if I can change this during the hackathon 💪

[pinin4fjords]

#175, combined with pinin4fjords/shinyngs#57, addresses this issue.