ECHTVAR_ANNO gives "...variants must be decomposed before running..."

[A97paupic]

Description of the bug

Hi again, I don't know for sure but I think that ECHTVAR_ANNO would like a vcf where multi-allelic sites has been split beforehand? Would greatly appreciate if someone could have a look.

Best Regards,
Paul

Command used and terminal output

ERROR ~ Error executing process > 'GENOMICMEDICINESWEDEN_SKIERFE:SKIERFE:SNV_ANNOTATION:ECHTVAR_ANNO (pr_001_001)'

Caused by:
  Process `GENOMICMEDICINESWEDEN_SKIERFE:SKIERFE:SNV_ANNOTATION:ECHTVAR_ANNO (pr_001_001)` terminated with an error exit status (101)

Command executed:

  echtvar anno -e gnomad.v3.1.2.echtvar.v2.zip -e multisample.zip pr_001_001.vcf.gz pr_001_001.anno.bcf.gz

  cat <<-END_VERSIONS > versions.yml
  "GENOMICMEDICINESWEDEN_SKIERFE:SKIERFE:SNV_ANNOTATION:ECHTVAR_ANNO":
      echtvar: $(echo $(echtvar -V) | sed 's/echtvar //' )
  END_VERSIONS

Command exit status:
  101

Command output:
  (empty)

Command error:
  fields: [Field { field: "AC", alias: "gnomad_ac", missing_value: -1, missing_string: "MISSING", zigzag: false, multiplier: 1, ftype: Integer, values_i: 0 }, Field { field: "AN", alias: "gnomad_an", missing_value: -1, missing_string: "MISSING", zigzag: false, multiplier: 1, ftype: Integer, values_i: 1 }, Field { field: "nhomalt", alias: "gnomad_nhomalt", missing_value: -1, missing_string: "MISSING", zigzag: false, multiplier: 1, ftype: Integer, values_i: 2 }, Field { field: "AF", alias: "gnomad_af", missing_value: -1, missing_string: "MISSING", zigzag: false, multiplier: 5000000, ftype: Float, values_i: 3 }, Field { field: "AC_popmax", alias: "gnomad_popmax_ac", missing_value: -1, missing_string: "MISSING", zigzag: false, multiplier: 1, ftype: Integer, values_i: 4 }, Field { field: "AN_popmax", alias: "gnomad_popmax_an", missing_value: -1, missing_string: "MISSING", zigzag: false, multiplier: 1, ftype: Integer, values_i: 5 }, Field { field: "nhomalt_popmax", alias: "gnomad_popmax_nhomalt", missing_value: -1, missing_string: "MISSING", zigzag: false, multiplier: 1, ftype: Integer, values_i: 6 }, Field { field: "AF_popmax", alias: "gnomad_popmax_af", missing_value: -1, missing_string: "MISSING", zigzag: false, multiplier: 5000000, ftype: Float, values_i: 7 }, Field { field: "AF_controls_and_biobanks", alias: "gnomad_controls_and_biobanks_af", missing_value: -1, missing_string: "MISSING", zigzag: false, multiplier: 5000000, ftype: Float, values_i: 8 }, Field { field: "nhomalt_controls_and_biobanks", alias: "gnomad_controls_and_biobanks_nhomalt", missing_value: -1, missing_string: "MISSING", zigzag: false, multiplier: 1, ftype: Integer, values_i: 9 }, Field { field: "FILTER", alias: "gnomad_filter", missing_value: 3, missing_string: "PASS", zigzag: false, multiplier: 1, ftype: Categorical, values_i: 10 }]
  fields: [Field { field: "AF", alias: "COHORT_AF", missing_value: -1, missing_string: "MISSING", zigzag: false, multiplier: 1000000, ftype: Float, values_i: 0 }, Field { field: "AC", alias: "COHORT_AC", missing_value: -1, missing_string: "MISSING", zigzag: false, multiplier: 1000000, ftype: Integer, values_i: 1 }]
  thread 'main' panicked at '[echtvar] variants must be decomposed before running. got variant with 3 alleles at 1:54720 ([[67, 84, 84, 84, 67, 84, 84, 84, 67, 84], [67], [67, 84, 84, 67, 84], [67, 84, 84, 67, 84, 84, 84, 67, 84]]). see: https://github.com/brentp/echtvar/wiki/decompose', /home/brentp/src/echtvar/src/lib/echtvar.rs:321:13
  note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

Work dir:
  /fs1/paul/LRS/240409/skierfe/work/cf/93162a4ffecdac99e3d8de8aee539c

Tip: view the complete command output by changing to the process work dir and entering the command `cat .command.out`

 -- Check '.nextflow.log' file for details

Relevant files

No response

System information

• 23.04.2
• HPC
• slurm
• singularity/3.8.0
• Linux

[fellen31]

Hi, sorry I've been away for a few days.

I think we might just change out the annotation workflow #23, but for the time being, yes, ECHTVAR_ANNO expects decomposed variants, which should be done in BCFTOOLS_NORM_SINGLESAMPLE.

The test profile is currently run without the SNV-annotation workflow because of the strict database requirements, and I might have broke BCFTOOLS_NORM_SINGLESAMPLE when updating the config. Could you share the .command.sh, .command.err and .command.log for BCFTOOLS_NORM_SINGLESAMPLE?

[A97paupic]

Sure, no probs. It runs fine but I think it lacks the flag to split multi-allelic sites(-m).

BCFTOOLS_NORM_SINGLESAMPLE.zip

[fellen31]

Thanks. The config looks broken at the moment, which causes BCFTOOLS_NORM_SINGLESAMPLE to run without -m.

Could you see if adding the below code to a custom config and supplying it to nextflow with -c works?

process {

    withName: '.*:SNV_ANNOTATION:BCFTOOLS_NORM_SINGLESAMPLE' {
        ext.args = "-m - --output-type b -w 10000"
    }

}

[A97paupic]

Yes, it sure did the trick! Thanks!

[fellen31]

Perfect. I'll fix this and work on adding the annotation workflow to the test profile, to avoid these kinds of mistakes in the future.

[fellen31]

Closed with #75.