Rnaseq: Locale error in step Picard MarkDuplicates

[santiagorevale]

Hi there!

I’m running version 1.1 of the pipeline using the Singularity’s image provided in here. When running Picard MarkDuplicates I’m getting the following error:

/opt/conda/envs/nf-core-rnaseq-1.1/bin/picard: line 5: warning: setlocale: LC_ALL: cannot change locale (en_US.UTF-8): No such file or directory

Is anybody else having the same issue? Is it something that should be fixed in the image, right? I tried running something like apt-get install locales but don’t image is read-only (don’t have much experience in Singularity, should I be able to bypass the read-only limitation?).

Thank you very much in advance for any help.

Cheers,
Santiago

[ewels]

Hi @santiagorevale,

Thanks for reporting this. It's the first time I've seen this error, which is strange - typically problems like this no longer happen when using Singularity. The fact that you're using v1.1 and I only released this two days ago makes me suspicious though (also the fact that I updated Picard in this release).

Have you tried running v1.0 of the pipeline? Do you get the same error? You should be able to run it by just adding -r 1.0 to your nextflow command.

Phil

[santiagorevale]

Hi @ewels,

I forgot to mention that besides the warning message, the pipeline keeps running. Sorry for that.

However, I've tried running it with version 1.0 and the same warning is happening. On the other hand, I came across an error on version 1.0 when running on Nextflow version 0.32.0 which I'm detailing below. Please, let me know if you want me to create a new issue with it and remove it from here.

Oct-08 09:05:11.573 [Actor Thread 42] ERROR nextflow.processor.TaskProcessor - Error executing process > 'star (Ganglion_WT_S2)'

Caused by:
  Not a valid memory value: null

Source block:
  prefix = reads[0].toString() - ~/(_R1)?(_trimmed)?(_val_1)?(\.fq)?(\.fastq)?(\.gz)?$/
  def avail_mem = task.memory == null ? '' : "--limitBAMsortRAM ${task.memory.toBytes() - 100000000}"
  seqCenter = params.seqCenter ? "--outSAMattrRGline ID:$prefix 'CN:$params.seqCenter'" : ''
  """
  STAR --genomeDir $index \\
      --sjdbGTFfile $gtf \\
      --readFilesIn $reads  \\
      --runThreadN ${task.cpus} \\
      --twopassMode Basic \\
      --outWigType bedGraph \\
      --outSAMtype BAM SortedByCoordinate $avail_mem \\
      --readFilesCommand zcat \\
      --runDirPerm All_RWX \\
      --outFileNamePrefix $prefix $seqCenter \\
  """

Tip: when you have fixed the problem you can continue the execution appending to the nextflow command line the option `-resume`
Oct-08 09:05:11.575 [Actor Thread 42] DEBUG nextflow.Session - Session aborted -- Cause: Process `null` script contains error(s)

[ewels]

Aha - if it keeps running then it's possible that we noticed the error but figured it wasn't causing any problems so just ignored it. I assume the results look ok?

The error with null memory is indeed specific to nextflow 0.32.0 and is fixed in the most recent release. See #95

Ok - so if everything is working as expected, then I might close this if that's ok 😁I don't think that a warning message with no consequence on the results is worth fighting with. If you think it's more severe than this then please let us know though.

Phil

[hekatonkhieres]

Hi @ewels ,

I noticed in your post above that you said it "no longer" happens. That means the "cannot change locale" issue happened before but was resolved somehow?

Can you share what the fix was for this? I am stuck using Nextflow v0.20.1 and I experience this issue now and then. I'd love to be able to solve it.

[ewels]

Hi @hekatonkhieres,

v0.20.1 is really really old. I’m kind of amazed that the pipeline runs at all, as we make use of several features that were released in later versions of nextflow. How come you’re locked to that version?

I don’t really want to get in to trying to fix this on such an old version of nextflow tbh.

Phil