GATK GQC always cast to int to prevent genmod from ignoring it

[alkc]

Description

fixes #170

Small change to prescore_sv.pl that always casts GATK GQC vals as ints.

Genmod expects an int and won't process GQC otherwise, which affects about 700 SVs per run where the caller penalty is not correctly set.

This update will enable the caller penalty for those on average 700 variants.

Risk assessment:

Fixing this bug can affect on average around 700 out of 2500 GATK-only per run, looking at the final SV results.

An analysis of confirmed CNVs only called by GATK used in our 2021 validation of the SV rank model confirms that this update is unlikely to affect the rank scores of confirmed variants.

These variants have GQC values > 2000, which means that the caller penalty would have been 0 even in absence of the bug this update fixes.

Type of change

• Patch

Checklist:

• [x] My code follows the style guidelines for NF pipelines at CMD
• [x] I have performed a self-review of my code
• [x] I have commented my code, particularly in hard-to-understand areas
• [x] My changes generate no new warnings (Keep an eye on .nextflow.log !)
• [x] I have updated the CHANGELOG
• [x] The latest commit in the master branch is tagged with the correct version number

Patch

• [x] Stub run completes without errors or new warnings
• [x] At least one other person has reviewed and approved my code

Instructions for the reviewers

How to test the changes

• [x] Compare the GQC vals for SV results from this branch with a live run and confirm that they are approximately equal (some slight variation is to occur)
• [x] Confirm that all GATK-only SVs are properly penalized based on their GQC val. A plot of GQC against the caller-penalty rank score component can be of help here.

Expected outcome

[Optional] Additional information

Review

Review performed by:

• Reviewer 1: Viktor, @ViktorHy
• ...

Testing performed by:

• Alex, @alkc

Post-merge

• [ ] The master branch has been tagged with the new version number.

[alkc]

Thanks. I'll run the tests for a wgs single/trio and merge if everything looks ok. Should be able to pull it off before the end of the week.

[alkc]

I checked the intermediate files for a live run and the corresponding sample run through this branch.

No obvious errors, the GQC values are converted to ints as they should. Running genmod score on the updated output results in no GATK Found value None errors for records with GQC defined.

[alkc]

Can confirm that the caller penalty for GATK work as it should now.