vg aument Bus error (core dumped)

[SimonaSecomandi]

1. What were you trying to do?

I was trying to performe the augmentation step before variant calling to call variants from a set of illumina WGS reads that I've aligned to my vg graph. I'm running the command with 32 threads and 256 GB RAM.

2. What did you want to happen?

Generate the augmentation file to use for variant calling

3. What actually happened?

/var/spool/slurm/d/job470186/slurm_script: line 15: 11547 Bus error (core dumped)

5. What data and command can the vg dev team use to make the problem happen?

Previous commands:

vg mod -t 32 -X 256 .pangenome.vg > pangenome_chopped.vg
vg index -t 32 -x pangenome_chopped.xg pangenome_chopped.vg
vg prune -t 32 -k 45 pangenome_chopped.vg > pangenome_chopped_pruned.vg
vg index -t 32 -b /tmp -p -g pangenome_chopped_pruned.gcsa pangenome_chopped_pruned.vg
vg map -t 32 -f WGS_forward.fastq.gz -f WGS_reverse.fastq.gz -x pangenome_chopped.xg -g pangenome_chopped_pruned.gcsa > WGS_aln.gam

Aumentation command that caused the bus error:

vg augment pangenome_chopped.vg WGS_aln.gam -A WGS_aug.gam > WGS_aug.vg

6. What does running vg version say?

vg version v1.30.0 "Carentino" Compiled with g++ (Ubuntu 9.3.0-17ubuntu1~20.04) 9.3.0 on Linux Linked against libstd++ 20200808 Built by anovak@octagon

Many thanks!

[adamnovak]

Unfortunately without the input files or the traceback it's difficult for us to help with this. If you run with the VG_FULL_TRACEBACK=1 environment variable, vg should spit out a traceback to standard error when it crashes that should implicate some part of the code as the source of the problem.

If you can post that, or the input files for the augment step, we might be able to identify and fix the issue.

All I can say right now is that either vg augment is wrong somehow, or 256 GB RAM isn't actually enough (and we're not checking for failed memory allocations somewhere when we should be).

[glennhickey]

Two ways to reduce vg augment's memory consumption:

• work on PackedGraph format (convert with vg convert -p or check the format with vg stats -F)
• specify a minimum coverage with -m to avoid adding read errors to the graph.

[SimonaSecomandi]

Many thanks for you answer!! I managed to generate the augmented graph on a 1 TB RAM node!