out_of_range error

[yuliamostovoy]

Hi,

I've had good results from running Translocator on a local server, and now I'm having trouble trying to run it through a WDL file on the google cloud infrastructure. With every input file I try, I get this in the log file:

Finalizing  ..
Number of denovo bp: 0
Number of total bp: 1328
terminate called after throwing an instance of 'std::out_of_range'
  what():  _Map_base::at
/cromwell_root/script: line 26:    16 Aborted                 (core dumped) translocator -m /cromwell_root/fc-4f147b38-30db-4e2b-8693-ce0d6c0a8f9d/results/ONTWholeGenome/SM-J9ECN/alignments/SM-J9ECN.bam -a /cromwell_root/broad-dsde-methods-long-reads/resources/references/chm13/chm13.draft_v1.0.fasta -v J9ECN.translocator.vcf --min_het_af 0.1 --global_remap

I've tested it on the same input files and parameters that successfully ran on the local server, and they generate that same error. I've checked to confirm that it's not running out of disk space or RAM. It's running from a docker image based on Ubuntu 16.04. Any help would be greatly appreciated!

[SHuang-Broad]

Hi @aquaskyline,

I've looked at the error message further and it looks like the error might be due to

• a non-number being passed to seqan
• some bug in seqan itself

Specially, the error message

terminate called after throwing an instance of 'std::out_of_range'
what(): _Map_base::at

Suggests that some unexpected keys are queried when using the map structure in seqan, hence the out_of_range exception, as is expected here.

So we are wondering if there's a way to turn on debug mode, and/or to get down to the offending alignments?

Thanks!

Steve

---

One other possible reason—though extremely unlikely—is that the T2T CHM13 v1.0 reference is used instead of the typical GRCh38 or HG19. So could it be a hard coded parameter that leads to a mismatch between Translocator and the data (new reference)?

[LeaveYeah]

Hi Steve,

One probable reason I can think of is that the bam and fasta files you use may have different chromosome names, e.g. one starting with "chr" and the other not. It would cause the out_of_range error as seqan cannot find the chromosome in the fasta.

Best,
Alex

[yuliamostovoy]

Hi Alex,

Thanks for the response! I wish it were that straightforward but the bam file was aligned to the exact same fasta file as we're inputting to Translocator. I just double-checked and both have chromosome names starting with 'chr'.

Thanks,
Yulia

[LeaveYeah]

Hi Yulia,

Well, that seems pretty mysterious. I suppose I could add some dubug info for you to pinpoint where the problematic read is. I am now preparing for a move to another city. Hopefully I can find time to do it next week.

Best,
Alex