nf: derive --organism from params.genome, gate ARDA on supported assemblies

[mikessh]

Closes #75.

@EgorGuga requested that the Nextflow module's nextflow.config:

1. run ARDA only when params.genome is GRCh38 or GRCm39, and
2. set --organism to human / mouse from the assembly.

The module's main.nf already honours both hooks this needs — task.ext.when (its when: block) and task.ext.args — so the change is entirely in the config, no module edit:

ext.when = { params.genome in ['GRCh38', 'GRCm39'] }
ext.args = {
    params.genome == 'GRCh38' ? '--organism human' :
    params.genome == 'GRCm39' ? '--organism mouse' : ''
}

params.genome is the iGenomes key nf-core sets from --genome, so ARDA follows whatever assembly the rest of the pipeline is already running on — nothing extra to configure. Any other genome (or none) skips ARDA and the pipeline still completes.

Verified end-to-end (not just parsed)

On aldan3 with real Nextflow (21.10.6), the real module + this config, and a fake arda on PATH that records its argv:

--genome	ARDA runs?	--organism received
GRCh38	✅	human
GRCm39	✅	mouse
GRCh37	skipped	—
(unset)	skipped	—

The skipped cases complete with no error; the two supported cases publish all four output files.

Docs updated (module README.md + pipeline_integration.rst): the genome→organism mapping, and that a static ext.args string overrides the genome-driven default. Sphinx -W build clean.