Add deepTools to get bigWig coverage files

[szilvajuhos]

bigWig is a compact format to get coverage information, and we can add it to IGV (https://github.com/igvteam/igv.js/wiki/Wig-Track) . With deepTools (https://deeptools.readthedocs.io/en/develop/index.html) we can easily calculate this file like

bamCoverage -bs 10 -v -b Preprocessing/Recalibrated/sample.recal.bam --ignoreDuplicates --numberOfProcessors 48 --outFileName coverageDir/sample.coverage.bigWig

It would be nice to have it among the results (and deepTools in the container, see conda install at: https://deeptools.readthedocs.io/en/develop/content/installation.html) . Maybe it can replace QualiMap and provide the same QC stuff.

[maxulysse]

Small issue, currently Manta is still in Python 2.7 with no plan to move to Python 3 yet: Illumina/manta#180
This raises conflict when trying to resolve conda environment:

UnsatisfiableError: The following specifications were found to be in conflict:
  - deeptools=3.3.1
  - manta=1.6.0

[maxulysse]

I think I'll more into it next year, when we start looking more into modules

[maxulysse]

UnsatisfiableError: The following specifications were found to be in conflict:
  - deeptools=3.3.1 -> matplotlib[version='>=3.0.0'] -> matplotlib-base==3.0.2=py37h167e16e_1002
  - deeptools=3.3.1 -> matplotlib[version='>=3.0.0'] -> python_abi=[build=*_cp37m] -> pypy[version='<0a0']
  - tiddit=2.7.1

Still conflict, so I'll try again when I'm finished with DSL 2

[szilvajuhos]

OK, if we are moving towards 1-process-1-container, can we have this included somehow? Otherwise I can look how to generate bigWig with other tools.

[maxulysse]

Yes, definitively that's the plan, I'm just trying to release Sarek 2.6 first, and then it's full on DSL 2

[FriederikeHanssen]

@szilvajuhos after which step do you want this output? We have now implemented splitting the fastq files -> mapping splits in parallel -> duplicateMarking with implicit merging. Would deepTools after duplicate marking be ok (in the case that duplicate marking is not skipped anyway)? Otherwise we have to do the extra merging step again or you get a bunch of file for each read pair. From the documentation I understand the deep tools can only process one BAM file as input and not a list

[szilvajuhos]

Huh, being honest, I have no faintest idea. It seemed like a good idea at that time, but I do not think it is urgent to add if at all.

[FriederikeHanssen]

Ok, well I was about to smash some old issues now and this was on the list. In that case for simplicity I will add it in after duplicate marking, because that is when the other QC metrics are computed for now.

[FriederikeHanssen]

Added by PR #479 It was quick to add so here we go. let me know what you think :)