ok another potential threat to patient safety occurs when you run the WES version of the pipeline due the use of fewer CNV calling methods (so just gatk cnv+ manta) this likely also affects WGS but if CNVnator catches them (they should be easy to find) they should at least come out... so maybe not a huge priority for everyone. this time I don't have an easy fix because it depends on your reference genome... if you don't specify the allosomes : chrX/chrY duplications will be missed in male patients due to the default methods of https://gatk.broadinstitute.org/hc/en-us/articles/30332057809179-PostprocessGermlineCNVCalls which treats all chrs as autosomes (aka diploid) unless otherwise specified [i.e. it drops CN = 2 on chrX even in males]. so my fix internally is to add this to my config withName: '.*CALL_SV_GERMLINECNVCALLER:GATK4_POSTPROCESSGERMLINECNVCALLS' { ext.args = '--allosomal-contig chrX --allosomal-contig chrY' } I tried to do a blanket --allosomal-contig X --allosomal-contig Y --allosomal-contig chrX --allosomal-contig chrY for all use-cases as a fix but it will error out when it matches back to the FASTA configs so it has to be custom set based on the reference genome

Great catch! We already set the default parameters for some processes based on the genome version like this one. What do you think about setting something similar for postprocessgermlinecnvcalls?

ah ok didn't see that yeah I guess that would work. we could add it here? https://github.com/nf-core/raredisease/blob/12eafba3db40d76ed2f7bd27414e169ce89d3e01/conf/modules/call_sv_germlinecnvcaller.config and pretty much copy that whole structure with the meta.sex. so like

withName: '.*CALL_SV_GERMLINECNVCALLER:GATK4_POSTPROCESSGERMLINECNVCALLS' {

ext.args = { meta.sex == 1

? params.genome == 'GRCh37'

? '--allosomal-contig X --allosomal-contig Y'

: '--allosomal-contig chrX --allosomal-contig chrY'

: '' }

}

Yupe! That sounds great!

ok ill test it internally and then make an issue + pull request

Perfect! Thanks, Benjamin!

ok finally got around to make a PR github.com/nf-core/raredisease/pull/966