lrp_4.0.16

Updated short read variant calling and joint calling pipelines to run at scale (#450)

Single-sample variant calling:

• Fixied critical typo in HaplotypeCaller disk spec.
• Fixed wdl-computed divide by zero error in SRFlowcell outputs.
• Updated MergeVCFs to have an option to name an output as a GVCF.
• Removed deprecated GC logging flags from GATK commands.
• Fixed issue in FastQC if no read qualities are in the file.
• Added missing annotation groups to ReblockGVCFs.
• HaplotypeCaller WDL now returns the reblocked GVCF.

Joint Calling:

• Added option to use gnarly genotyper
• Added het inputs to joint genotyping
• Fixed java memory allocation in joint genotyping to be based on memory of the VM, not hard-coded
• Fixed name of outdir in ConvertToZarrStore to be correct for this workflow.
• Updated the zarr conversion to use parallel Dask processes and to log to stdout.
• Upped default zarr conversion memory to 32gb.

VETS:

• Added stack trace logging for errors in ExtractVariantAnnotations,
  TrainVariantAnnotationsModel, and ScoreVariantAnnotations.
• Removed HAPCOMP, HAPDOM, and HEC from default annotations for SNP and INDEL VETS filtration on joint-called VCF files. Need to do more testing / debugging to include these in joint calling.

Misc:

• Fixed the name of the workflow in ExpandedDrugResistanceMarkerAggregation to reflect the actual name.
• Miscellaneous updates for debugging