Don't run samtools mpileup when only tools mpileup is selected

[FriederikeHanssen]

Description of the bug

currently samtools mielup and bcftools mpileup are both run when tools mpileup is selected, however only run bcftools then + fix the output docs if this is done

Command used and terminal output

No response

Relevant files

No response

System information

No response

[asp8200]

This whole things seems very delicate, and we should probably be very careful when changing the current setup.

The rather puzzling usage of both samtools mpileup and bcftools mpileup seems to be due to problems with getting the mpileup-subworkflow to play nicely with controllfreec :

#832 (comment)

[FriederikeHanssen]

so basically samtolls mpileup should be run for controlfreec. But when people select mpileup in tools, bcftools should be used

[asp8200]

so basically samtolls mpileup should be run for controlfreec. But when people select mpileup in tools, bcftools should be used

That sounds like one way (and probably the right way) of fixing this.

However, 'i'd say that Amin's immediate problem is that when Sarek runs with intervals no vcf-file (just mpileup-file) is published, and that happens even though BAM_VARIANT_CALLING_MPILEUP:GATK4_MERGEVCFS is called :

[bb/93ce90] process > NFCORE_SAREK:SAREK:BAM_VARIANT_CALLING_GERMLINE_ALL:BAM_VARIANT_CALLING_MPILEUP:BCFTOOLS_MPILEUP (testT)            [100%] 4 of 4 ✔
[ab/e1ab50] process > NFCORE_SAREK:SAREK:BAM_VARIANT_CALLING_GERMLINE_ALL:BAM_VARIANT_CALLING_MPILEUP:SAMTOOLS_MPILEUP (testT)            [100%] 4 of 4 ✔
[b7/18fc91] process > NFCORE_SAREK:SAREK:BAM_VARIANT_CALLING_GERMLINE_ALL:BAM_VARIANT_CALLING_MPILEUP:CAT_MPILEUP (testT)                 [100%] 2 of 2 ✔
[11/a1d7aa] process > NFCORE_SAREK:SAREK:BAM_VARIANT_CALLING_GERMLINE_ALL:BAM_VARIANT_CALLING_MPILEUP:GATK4_MERGEVCFS (testT)             [100%] 2 of 2 ✔

A quick-fix for Amin might be to add a config for publishing from BAM_VARIANT_CALLING_MPILEUP:GATK4_MERGEVCFS.

[FriederikeHanssen]

yep that sounds very reasonable. Do you have time to guide him through it?

[asp8200]

yep that sounds very reasonable. Do you have time to guide him through it?

I'll try as best I can

[asp8200]

So in conf/modules/mpileup.config, I added

withName: 'NFCORE_SAREK:SAREK:BAM_VARIANT_CALLING_GERMLINE_ALL:BAM_VARIANT_CALLING_MPILEUP:GATK4_MERGEVCFS' {
    ext.when   = { (meta.num_intervals > 1) && params.tools && params.tools.split(',').contains('mpileup') }
    publishDir = [
        mode: params.publish_dir_mode,
        path: { "${params.outdir}/variant_calling/mpileup/${meta.id}/" },
        pattern: "*{vcf.gz,vcf.gz.tbi}"
    ]
}

And, with that, I get testN.normal.mpileup.gz, testN.vcf.gz and testN.vcf.gz.tbi in results/variant_calling/mpileup/testN/ when running

nextflow run main.nf -profile test,targeted,docker --input ./tests/csv/3.0/mapped_joint_bam.csv --tools mpileup --step variant_calling --outdir results

Disclaimer: This quick-fix hasn't been tested with controllfreec.

@ Amin : Let me know if that works for you.