[Bug] DNAscope gnomad rule failing due to memory

[mathiasbio]

Description

A couple of WGS cases have now been unable to complete in the DNAscope gnomad rule which creates a BAF file to be used as input in GENS. See deviation: https://github.com/Clinical-Genomics/Deviations/issues/631

The memory issues seem to appear after entry into the HLA-region in chromosome 6, with a warning like this:
2024-05-07/05:43:07.323620 HCAssembler WARN :skip selectHaplotype 6:32492321-32492798 due to big read number 44 and big hap number 34117160

But it's strange that the cases keep increasing memory even after it has finished processing this region, and that it's so sample dependent, where these problematic samples can't finish even with 400G mem, whereas other samples finish entirely without having used more than around 30G.

How to reproduce

No response

Expected behaviour

No response

Anything else?

No response

Pipeline version

No response

[mathiasbio]

It seems after my own manual tests that after removing the variants in the gnomad VCF that exists in
chr6 32393697-32787282 (HLA region)

The DNAscope gnomad command finishes, whereas running with the original one fails. I believe this region is too small to matter in the GENS visualisation that we could replace the current gnomad VCF with this HLA filtered one in production, and hopefully that fixes this issue with memory fails.

But does this require us to make a new version of balsamic?

[mathiasbio]

Replaced by user-story #1447