Add more entry points to the workflow

[ramprasadn]

Description of feature

It would be nice to have entry points for different parts of the pipeline ex. snv/sv annotation, mitochondrial analysis.

[ramprasadn]

break this down into smaller issues

1. start from duplicate marked bam
2. start from variant called vcfs

[fa2k]

I've made a draft version of this. This is just to have something concrete to look at - I don't think it's necessarily the right way, as I don't know the pipeline that well.

• Add an input_type parameter to the workflow, which can either be reads for FASTQ (default) or alignments for BAM. In the future there could also be another value for VCF+BAM.
• Add a column bam to the sample sheet. The CHECK_INPUT process is used to get the BAM and BAI files based on the sample sheet.
• Add a test_bam config to test it

Incomplete:

• Modify SAMPLESHEET_CHECK and its script
• Test bam file and samplesheet are not uploaded to test data repository

Problem:

• BWAMEM2_MEM_MT crashes ' [E::bwa_set_rg] the read group line is not started with @rg'. (I don't know yet if this is a serious problem with the approach, or a trivial fix)

See changes in my fork fa2k@ff78904

[fa2k]

I've made it run with both the existing test and a new test for bam input (and cleaned up a bit).

The test outputs are not identical, but I've checked two vcf files:

annotate_snv/justhusky_rohann_vcfanno_filter_vep.vcf: Identical up to different timestamps in headers
annotate_sv/justhusky_svdbquery_vep.vcf: Unknown differences

check_samplesheet.py I made a polymorphic RowChecker - it's a bit strange and we can consider alternatives. Overall, here's the changes compared to the dev branch:

dev...fa2k:raredisease:multiple-entry-points

The test_bam profile requires an override sample sheet, and needs bam file to exist locally.

[fa2k]

I have updated to integrate the upstream changes from dev. Will make a pull request.

[Jakob37]

Any updates on this feature? Starting from bam-files would be extremely useful.

[fa2k]

Any updates on this feature? Starting from bam-files would be extremely useful.

Sorry for the late reply. As far as I know, there is no work that has been started. I'm even unlikely to start in the near future. I still need it, and will start eventually if nobody else takes it.

[maxulysse]

I'd happily share the logic with have in Sarek for this, and really we should converge on more subworkflows and bits of code for this kind of things