GATK-SV expects all genotypes to be biallelic, even on sex chromosomes, for legacy reasons.
based on this, there is some division during the GATK-SV annotation process which does a int/int division, expecting an int result.
GATK's gCNV doesn't use the same GT representations, so by pushing gCNV results through the GATK-SV pipeline, we're running into issues that the main GATK-SV team is not encountering.
This is fixable at source, by making script/workflow changes, releasing those changes, and releasing a new set of images with those updated scripts baked in, but this is a substantial distraction to solve an issue they don't run into.
This script takes the image version we are already using, copies a new version of this problem script into the image, and keeps all other content the same. The intention is that we can use this image in the gCNV pipeline to overcome this issue in the short term, and revert back once a fix is available at source.
The one functional change in this script is the addition of a round call when grabbing the Male and Female ANs. There are a couple of additional changes to suppress linting errors.
See https://github.com/broadinstitute/gatk-sv/pull/661 for context
This script takes the image version we are already using, copies a new version of this problem script into the image, and keeps all other content the same. The intention is that we can use this image in the gCNV pipeline to overcome this issue in the short term, and revert back once a fix is available at source.
The one functional change in this script is the addition of a
roundcall when grabbing the Male and Female ANs. There are a couple of additional changes to suppress linting errors.