Difficulties getting the pipeline run

[ktrns]

Hi guys!

We will be sequencing several hundred virus strains, and I am looking for a pipeline to do de-novo assembly and variant calling.

Since I have been working with nfcore @drpatelh in the past, I would like to try out this pipeline. The error message, I am currently getting is:

Error executing process > 'MINIA (L50982_v1_S1_ampfree)'

Caused by:
  Process `MINIA (L50982_v1_S1_ampfree)` terminated with an error exit status (1)

Command executed:

  echo "L50982_v1_S1_ampfree.viral_1.fastq.gz
  L50982_v1_S1_ampfree.viral_2.fastq.gz" > input_files.txt
      minia \
          -kmer-size 31 \
          -abundance-min 20 \
          -nb-cores 6 \
          -in input_files.txt \
          -out L50982_v1_S1_ampfree.k31.a20
      mv L50982_v1_S1_ampfree.k31.a20.contigs.fa L50982_v1_S1_ampfree.k31.scaffolds.fa

Command exit status:
  1

and further down in the .command.log:

  EXCEPTION: error opening file: L50982_v1_S1_ampfree.k31.a20.contigs.fa (No such file or directory)

The way I run the pipeline, is:

nextflow run nf-core/viralrecon --input $base/samplesheet.S1S2.csv --protocol 'amplicon' --amplicon_bed $ampliconBed --amplicon_fasta $ampliconFasta --genome 'MN908947.3' --igenomes_ignore --kraken2_db $kraken2Db --outdir $base -with-singularity $base/viralrecon_dev.sif -c dcgc.config -r dev

Would you be able to help me out? Best wishes Katrin

[drpatelh]

Hi @ktrns ! Hope you are and your loved ones are keeping well :)

The minia assembler was just added to the pipeline a couple of days ago so we havent really stress tested it much. By default, the pipeline will run all assembly tools so maybe you can try re-running with --assemblers spades,metaspades,unicycler -resume.

Command looks great other than the fact that the pipeline isnt using AWS iGenomes so you can remove the --igenomes_ignore flag.

I would like to know why minia is failing though. Can you cd into the work directory for the process that failed and send me the output of ls -la please? May be easier to discuss this in real-time on the #viralrecon channel on nf-core Slack.

[drpatelh]

This was partly fixed by using the human Kraken2 database as made available here and not a Kraken2 database built using the --standard flag as this will filter out most of the reads including the viral ones that we are interested in. Other issues have been reported on Slack but these relate to not enough viral reads being available to go through some of the assembly steps.