Background

Steven ran Bismark to align 24 C. gigas samples but ran into issues using ckpt on klone because of the time limit. He eventually was able to complete the analysis here: https://sr320.github.io/tumbling-oysters/posts/40-HawsOA/

He asked if I could get it to run with the nf-core pipeline methylseq

Results

After correctly modifying the process section of the .config file to match his alignment parameters, I was able to successfully reproduce his results in 16.5 hours:

Methods

Modifying .config file to include Bismark parameters

I added the following chunck to the .config file. Full .config file is on Klone at /gscratch/srlab/strigg/bin/uw_hyak_srlab.config and here.

  withName: BISMARK_ALIGN {
        ext.args = {
          [
            '--score_min L,0,-0.8',
	    '--non_directional'
          ].join(' ')
        }
        ext.prefix = { "${meta.id}" }
    }
   
   withName: BISMARK_METHYLATIONEXTRACTOR {
        ext.args = {
	  [
	   '--merge_non_CpG',
	   '--comprehensive',
	   '--multicore 48',
	   '--buffer_size 75%'
	  ].join(' ')
	}
	ext.prefix = { "${meta.id}" }
    }

   withName: BISMARK_COVERAGE2CYTOSINE {
        ext.args = {
          [
            '--merge_CpG',
            '--zero_based'
          ].join(' ')
        }
        ext.prefix = { "${meta.id}" }
    }
}

Running the methylseq pipeline on Klone

I ran the following code to reprocess the data:

#start screen session 
screen -S methylseq

#start interactive node
salloc -A srlab -p cpu-g2-mem2x -N 1 -c 1 --mem=16GB --time=72:00:00

#activate nextflow conda environment
mamba activate nextflow

#run methylseq pipeline
nextflow run nf-core/methylseq \
-c /gscratch/srlab/strigg/bin/uw_hyak_srlab.config \
--input /gscratch/scrubbed/strigg/analyses/20250401_methylseq/samplesheet.csv \
--outdir /gscratch/scrubbed/strigg/analyses/20250415_methylseq \
--fasta /gscratch/srlab/sr320/github/project-oyster-oa/data/Haws-11/GCF_963853765.1_xbMagGiga1.1_genomic.fa \
-resume \
-with-report nf_report.html \
-with-trace \
-with-timeline nf_timeline.html \
--skip_trimming \
--nomeseq

Other attempts that FAILED

  1. In this attempt, I was using the default alignment parameters (e.g. score_min L,0,-0.2) so the %aligned was much lower than Steven saw.

  2. In this attempt, I had the incorrect syntax in the .config file

  3. In this attempt, I still had the incorrect syntax in the .config file. See slack thread here: https://nfcore.slack.com/archives/CP3RJSMF0/p1743520726837279