DSL2 Planning

[jfy133]

Whelp, I guess we have to tackle the bull.

First thoughts, we need to work out what nf-core/modules we would need, check what are already available and what we would have to write.

List based on currently used tools in eager DSL1:

• [x] adapterremoval
• [ ] adapterremovalfixprefix (although honestly would rather drop this entirely
• [x] bwa
  • [x] aln (+ sampe/samse)
  • [x] mem
  • [x] index
• [ ] picard [Note: all in GATK4? Drop for GATK4 instead?]
  • [x] CreateSequenceDictionary
  • [ ] AddOrReplaceReadGroups
  • [x] MarkDuplicates
• [x] samtools
  • [x] faidx
  • [x] fastq
  • [x] index
  • [x] merge
  • [x] flagstat
  • [x] view
  • [x] sort
  • [x] #925
  • [x] mpileup
• [x] fastqc
• [x] dedup
• [x] angsd
  • [x] doCounts
  • [x] contamination
• [ ] circularmapper
  • [ ] circulargenerator
  • [ ] circularmapper
• [x] gatk4 HaplotypeCaller
• [ ] gatk3.5
  • [ ] UnifiedGenotyper
  • [ ] RealignerTargetCreator
  • [ ] IndelRealigner
  • [ ] UnifiedGenotyper
• [x] qualimap
  • [x] bamqc
• [ ] vcf2genome no one is using it..
• [x] damageprofiler
• [x] multiqc
• [ ] pmdtools
• [x] bedtools
  • [x] #779
• [x] sequencetools
• [x] preseq
  • [x] c_curve
  • [x] lc_extrap(?) <- requested
• [x] fastp
• [x] bamutil
  • [x] trimBam
• [x] mtnucratio
• [x] kraken2
• [x] freebayes
• [x] sexdeterrmine (+ samtools depth)
• [ ] multivcfanalyzer
• [x] hops
  • [x] MALT
  • [x] maltExtract + Posprocessing
• [x] bowtie2
  • [x] build
  • [x] align
• [x] eigenstratdatabasetools
  • [x] eingenstrat_snp_coverage (+ parse_snp_cov.py)
• [ ] mapdamage
• [x] bbmap
  • [x] bbduk.sh
• [x] cat (fastq lane merge)
• [ ] hostremoval_input_fastq
• [x] endorspy
• [ ] filter_bam_fragment_length.py
• [ ] print_x_contamination.py
• [ ] kraken_parse.py
• [ ] mege_kraken_res.py
• [x] bcftools
  • [x] stats

[jfy133]

[morien]

Just an end user here, but I'm hoping someone can tell me what the timeline for migration to DSL2 is. Trying to get eager running on my linux server and nextflow DSL1 conflicts with my java 18+, which I use for lots of other programs.

[jfy133]

We can't really say at the moment. 80-90% is implemented (the checklist above isn't up to date) or closed to being merged, but we will then have to go through a lot of testing.

I would say sometime first half of next year may be realistic?

But in the meantime - is conda not available on you cluster to install an earlier version of Java and nextflow (the latter you have to do anyway), for example?

[morien]

containerizing using conda proved difficult on our system, so we ended up using singularity and setting java default to 17 (rather than latest), which worked very smoothly. java default can be changed between v17 and latest, as needed, when not running eager. looking forward to support for nextflow 23.x in the meantime

[jfy133]

Glad to hear!