Error in DMR step

[npatel-ah]

Hello,

Thank you for the great pipeline. I was testing the latest version 2.0.0 with the GIAB data and got it following error. Not sure if it's due to using --all_context for human sample.

ERROR ~ Error executing process > 'NFCORE_METHYLONG:METHYLONG:DOWNSTREAM:DSS_HAPLOTYPE_LEVEL:MODKIT_PILEUP_HAPLOTYPE_LEVEL (HG003)'

Caused by:
  Process `NFCORE_METHYLONG:METHYLONG:DOWNSTREAM:DSS_HAPLOTYPE_LEVEL:MODKIT_PILEUP_HAPLOTYPE_LEVEL (HG003)` terminated with an error exit status (1)


Command executed:

  modkit \
      pileup \
      --motif CHH 0 --motif CHG 0 --motif CG 0  --combine-strands --partition-tag HP \
      --threads 12 \
      --prefix HG003 \
      --ref GCA_000001405.15_GRCh38_no_alt_analysis_set.fa \
       \
      HG003_haplotagged.bam \
      HG003.tmp

  if test -d HG003.tmp; then
      for file in HG003.tmp/*; do
          if test -f $file; then
              mv $file $(basename $file)
          fi
      done
  else
      mv HG003.tmp HG003.bed
  fi

  cat <<-END_VERSIONS > versions.yml
  "NFCORE_METHYLONG:METHYLONG:DOWNSTREAM:DSS_HAPLOTYPE_LEVEL:MODKIT_PILEUP_HAPLOTYPE_LEVEL":
      modkit: $( modkit --version | sed 's/mod_kit //' )
  END_VERSIONS

Command exit status:
  1

Command output:
  (empty)

Command error:
  Unable to find image 'quay.io/biocontainers/ont-modkit:0.5.0--hcdda2d0_2' locally
  0.5.0--hcdda2d0_2: Pulling from biocontainers/ont-modkit
  0cacab098358: Already exists
  bd9ddc54bea9: Already exists
  48fb8fbc4fc5: Pulling fs layer
  48fb8fbc4fc5: Verifying Checksum
  48fb8fbc4fc5: Download complete
  48fb8fbc4fc5: Pull complete
  Digest: sha256:40eb7c3ff0bef3d83ebf0b86bf2fb02ae4fb1975d363d1ae0c7d92a17954fc4b
  Status: Downloaded newer image for quay.io/biocontainers/ont-modkit:0.5.0--hcdda2d0_2
  > calculated chunk size: 18, interval size 100000, processing 1800000 positions concurrently
  > creating HG003.tmp
  > Error! cannot combine strands with a motif that is not a palindrome

Work dir:
  /mnt/PB/work/97/fa96db524feb34cdfccd97b21d5255

Container:
  quay.io/biocontainers/ont-modkit:0.5.0--hcdda2d0_2

Tip: you can try to figure out what's wrong by changing to the process work dir and showing the script file named `.command.sh`

 -- Check '.nextflow.log' file for details
ERROR ~ Pipeline failed. Please refer to troubleshooting docs: https://nf-co.re/docs/usage/troubleshooting

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

Here is my command
nextflow run nf-core/methylong --input ./samplesheet.csv --outdir ./results -profile docker --bedgraph --all_contexts --dmr_a dad --dmr_b son -r 2.0.0

Here is the samplesheet

group,sample,path,ref,method
dad,HG003,/data/PB/m84039_241002_000337_s3.hifi_reads.bc2020.bam,/data/PB/Minimap2/GCA_000001405.15_GRCh38_no_alt_analysis_set.fa,pacbio
son,HG002-rep1,/mnt/PB/m84039_241001_220042_s2.hifi_reads.bc2018.bam,/data/PB/Minimap2/GCA_000001405.15_GRCh38_no_alt_analysis_set.fa,pacbio
son,HG002-rep2,/mnt/PB/m84039_241002_040926_s1.hifi_reads.bc2024.bam,/data/PB/Minimap2/GCA_000001405.15_GRCh38_no_alt_analysis_set.fa,pacbio

Thank you.

[jkh00]

Hi,

Thanks for trying out the pipeline! :)

The problem lies in using --all-context and having --combine-strands hard-coded in modkit pileup.
--combine-strands only work for CG context. I am not sure why it was there in the first place, it should be fine to not include this in the script, correct me if I overlooked some issues @YiJin-Xiong. I created a patch for this fix (patch-2.0.1), could you try it out and let me know if it works? :)

Another issue I noticed here is that, the --dmr_population_scale parameter should be called explicitly here if you were intended to run population scale dmr analysis or else it doesnt run the dmr population workflow.

for example:
nextflow run nf-core/methylong --input ./samplesheet.csv --outdir ./results -profile docker --bedgraph --all_contexts --dmr_a dad --dmr_b son --dmr_population_scale -r patch-2.0.1

[YiJin-Xiong]

Thank you so much for the quick fix, really appreciated! @jkh00
Yes, you're absolutely right.
I originally used DSS as the default tool for population-scale DMR analysis, so I included the --combine-strands parameter for CG context analysis. It seems I forgot to remove it after adding --all-context, really sorry for that.

I've tested the patch-2.0.1, and made some minor changes.

[YiJin-Xiong]

Thank you for raising the issue! @npatel-ah We'll fix it as soon as possible.
Additionally, if you want to perform population scale dmr analysis, be sure to specify the --dmr-population-scale parameter.

[npatel-ah]

Thank you both for quickly looking into it. I will look forward to testing the updated version.