hey i have a lot of SRA to assemble some of them...
# bacass
j
hey i have a lot of SRA to assemble some of them don't survive the
skewer step
with the following error message:
Copy code
Error executing process > 'NFCORE_BACASS:BACASS:SKEWER (sample_01)'

Caused by:
  Process `NFCORE_BACASS:BACASS:SKEWER (sample_01)` terminated with an error exit status (1)

Command executed:

  # loop over readunits in pairs per sample
  pairno=0
  echo "sample_01_1.fastq.gz sample_01_2.fastq.gz" | xargs -n2 | while read fq1 fq2; do
      skewer -m pe -q 3 -n --quiet -t 12 $fq1 $fq2;
  done
  
  # gzip, because skewer's -z returns an error
  gzip *.fastq
  
  cat $(ls *trimmed-pair1.fastq.gz | sort) >> sample_01_trm-cmb.R1.fastq.gz
  cat $(ls *trimmed-pair2.fastq.gz | sort) >> sample_01_trm-cmb.R2.fastq.gz
  
  echo $(skewer --version 2>&1) | sed 's/^.*skewer version: //; s/ .*//' > skewer.version.txt

Command exit status:
  1

Command output:
  (empty)

Command error:
  WARNING: Skipping mount /var/singularity/mnt/session/etc/resolv.conf [files]: /etc/resolv.conf doesn't exist in container
  Error: the FASTQ quality formats of input files are different
is this a container issue:
WARNING: Skipping mount /var/singularity/mnt/session/etc/resolv.conf [files]: /etc/resolv.conf doesn't exist in container
or a skewer problem:
Error: the FASTQ quality formats of input files are different
... the pipeline does not complete from here. any idea how to fix this? thank you!
d
WARNING: Skipping mount /var/singularity/mnt/session/etc/resolv.conf [files]: /etc/resolv.conf doesn't exist in container
Is nothing to worry about.
Error: the FASTQ quality formats of input files are different
Indeed, this is the relevant part I expect. I think this is an input file problem. The arriving files are not as expected. Can you confirm that those are indeed fastq, and from same format (e.g. novaseq has binned quality scores, MiSeq not)?
j
will investigate - thank you for your response!
i tried to come up with some filtering - as shown in the snippet. i thought if biopython passes it, it should be okay but i also tried it in a more specific way -> but all the troublesome fastq files passed. to get around this, i wrote a super small pipeline, that basically tries skewer for all the fastq and only if it succeeds its put to the samplesheet.csv that i feed the bacass pipeline. this circumvents the issue - but its sad i could not figure out, what the actual issue was...
i put the helper pipeline here: https://github.com/plaquette/write_bacass_input_csv in case it helps someone at some point. [its rather quick and dirty than polished...]
d
Thanks! So skewer seems to be the problem. It might be also possible to use params for skewer that will allow the processing (but I havent tried much with it) and use a separate config which modifies skewer settings via https://github.com/nf-core/bacass/blob/959967364c7c0105b5b271acf0441fa9290e0d4c/conf/modules.config#L30-L34-
j
yes - it seems skewer has some sort of encoding issue (at least that's what i suspect since my data is collected from many places and might not share all characterstics) for my data its about 5-10% of the genomes that fail due to it ... which is okay.