Add IDR analysis

[drpatelh]

Would be nice to add the IDR analysis that is currently carried out by the ENCODE pipelines:
https://github.com/kundajelab/atac-seq-pipeline
https://docs.google.com/document/d/1f0Cm4vRyDQDu0bMehHD7P7KOMxTOP-HiNoIvL1VcBt8/edit

Another implementation in NF:
https://github.com/DoaneAS/atacflow

ENCODE ATAC-seq Guidelines:
https://www.encodeproject.org/atac-seq/

Amongst other things, this will probably also involve creating pseudo-replicates by merging and equally sampling alignments across replicates.

See: nf-core/atacseq#36

[JoseEspinosa]

This one could be also relevant for the implementation of the IDR analysis specifically for ChIP-seq analysis, implemented in NF:
https://github.com/bioinfo-pf-curie/chip-seq

[drpatelh]

We would need to think about how we define "replicates" here to use that right? Now that we have removed the replicate column from the samplesheet. Maybe we can have an optional column for replicates in the samplesheet for this kind of thing? 🤔

[JoseEspinosa]

Yes, that could be an option, another one could be to have a separate file with the information about replicates... In the case we want to keep the samplesheet standard across pipelines

[drpatelh]

I did think about having a separate metadata file but why?! Keep it simple and have one would be my preference.

[JoseEspinosa]

Yes, I agree two metadata files means increasing at least 2x the possibility of making mistakes 😝