Implement GATK-gCNV tool for CNV detection

[egenomics]

Description of feature

Hi,
It would be awesome if GATK-gCNV new tool for CNV detection in WES, targeted panels and WGS could be implemented in the pipeline.
The original paper is here: https://www.nature.com/articles/s41588-023-01449-0

[ramprasadn]

Hi @egenomics! Correct me if I am wrong, but this looks like a feature we already implemented in release 1.1.0. Relevant PR here #362.

[ramprasadn]

Closing this issue for now, but feel free to open it back up if you think this feature is lacking. 😄

[a113n]

In release 2.2.0, seems like the outputs from GATK CNV segment calls are not merged in a family-based VCF first:
`### name: 'NFCORE_RAREDISEASE:RAREDISEASE:CALL_STRUCTURAL_VARIANTS:SVDB_MERGE (fam_1)'

container: 'quay.io/biocontainers/mulled-v2-c8daa8f9d69d3c5a1a4ff08283a166c18edb0000:511069f65a53621c5503e5cfee319aa3c735abfa-0'

###BASH output
drwxr-xr-x 2 allen allen 4096 Oct 12 12:05 ./
drwxr-xr-x 4 allen allen 4096 Oct 5 11:24 ../
-rw-r--r-- 1 allen allen 0 Oct 5 11:24 .command.begin
-rw-r--r-- 1 allen allen 0 Oct 5 11:24 .command.err
-rw-r--r-- 1 allen allen 0 Oct 5 11:24 .command.log
-rw-r--r-- 1 allen allen 0 Oct 5 11:24 .command.out
-rw-r--r-- 1 allen allen 12434 Oct 5 11:24 .command.run
-rw-r--r-- 1 allen allen 602 Oct 12 11:58 .command.sh
-rw-r--r-- 1 allen allen 260 Oct 5 11:24 .command.trace
-rw-r--r-- 1 allen allen 1 Oct 5 11:24 .exitcode
lrwxrwxrwx 1 allen allen 101 Oct 5 11:24 sample1_gatkcnv_segments_refiltered.vcf.gz -> /mnt/data/wgs/work/e7/22f995203d96241c7692dc821f2c85/sample1_gatkcnv_segments_refiltered.vcf.gz
lrwxrwxrwx 1 allen allen 101 Oct 5 11:24 sample2_gatkcnv_segments_refiltered.vcf.gz -> /mnt/data/wgs/work/11/fc0c2445409f85f6a9f6b559685882/sample2_gatkcnv_segments_refiltered.vcf.gz
lrwxrwxrwx 1 allen allen 101 Oct 5 11:24 sample3_gatkcnv_segments_refiltered.vcf.gz -> /mnt/data/wgs/work/10/626e9f99ae2a6f70cc92b3a0d08e02/sample3_gatkcnv_segments_refiltered.vcf.gz
lrwxrwxrwx 1 allen allen 74 Oct 5 11:24 fam_1_cnvnator.vcf.gz -> /mnt/data/wgs/work/7d/6c9f89a9b60391bf328f71f8337819/fam_1_cnvnator.vcf.gz
lrwxrwxrwx 1 allen allen 82 Oct 5 11:24 fam_1_manta.diploid_sv.vcf.gz -> /mnt/data/wgs/work/fe/8bfb6f1fda184eccc08f79c9937cce/fam_1_manta.diploid_sv.vcf.gz
-rw-r--r-- 1 allen allen 0 Oct 12 12:05 fam_1_sv.vcf.gz
lrwxrwxrwx 1 allen allen 72 Oct 5 11:24 fam_1_tiddit.vcf.gz -> /mnt/data/wgs/work/fe/d528ed5cbaf1941ab1bafcb24e92e4/fam_1_tiddit.vcf.gz
-rw-r--r-- 1 allen allen 108 Oct 12 12:05 versions.yml
`
Expected sample1_gatkcnv_segments_refiltered.vcf.gz, sample2_gatkcnv_segments_refiltered.vcf.gz sample3_gatkcnv_segments_refiltered.vcf.gz are merged into fam_1_gcnvcaller.vcf.gz prior entering the NFCORE_RAREDISEASE:RAREDISEASE:CALL_STRUCTURAL_VARIANTS:SVDB_MERGE workflow.

Content in .command.sh may give insight on why the output fam_1_sv.vcf.gz is empty:
`
#.command.sh
#!/bin/bash -euo pipefail
svdb
--merge
--pass_only --same_order
--priority tiddit,manta,gcnvcaller,cnvnator
--vcf fam_1_tiddit.vcf.gz:tiddit fam_1_manta.diploid_sv.vcf.gz:manta sample1_gatkcnv_segments_refiltered.vcf.gz:gcnvcaller sample2_gatkcnv_segments_refiltered.vcf.gz:cnvnator sample3_gatkcnv_segments_refiltered.vcf.gz:null fam_1_cnvnator.vcf.gz:null
> fam_1_sv.vcf
bgzip fam_1_sv.vcf

cat <<-END_VERSIONS > versions.yml
"NFCORE_RAREDISEASE:RAREDISEASE:CALL_STRUCTURAL_VARIANTS:SVDB_MERGE":
svdb: $( echo $(svdb) | head -1 | sed 's/usage: SVDB-([0-9].[0-9].[0-9]).*/\1/' )
samtools: $(echo $(samtools --version 2>&1) | sed 's/^.samtools //; s/Using.$//')
END_VERSIONS`

Note that in the line with "sample3_gatkcnv_segments_refiltered.vcf.gz:null fam_1_cnvnator.vcf.gz:null", two files were labelled as null, potentially due to array overflow of unexpected number of input VCF files.

[a113n]

@ramprasadn Should the GATK SV workflow include a SVDB merge step as well (Similar to line 39 of the cnvnator workflow https://github.com/nf-core/raredisease/blob/fa61a657257a14a6433e3d751c577ab4c9d2eda4/subworkflows/local/variant_calling/call_sv_cnvnator.nf)?

[jemten]

Tracked in issue #634