Rnaseq: Singularity issue

[aubinthomas]

Hello,

I'm trying to use the pipeline through singularity. I had a look on other same issues but no way to see it working.

The commands I wrote are :

singularity pull --name nfcore-rnaseq.img docker://nfcore/rnaseq
./nextflow run nf-core/rnaseq -with-singularity nfcore-rnaseq.img -profile singularity --genome GRCh37 --reads '*{1,2}.fastq'

as output I have :

N E X T F L O W  ~  version 19.04.1
Launching `nf-core/rnaseq` [nasty_rubens] - revision: 37f260d360 [master]
nf-core/rnaseq v1.3
----------------------------------------------------
Pipeline Release  : master
Run Name          : nasty_rubens
Reads             : *{1,2}.fastq
Data Type         : Paired-End
Genome            : GRCh37
Strandedness      : None
Trimming          : 5'R1: 0 / 5'R2: 0 / 3'R1: 0 / 3'R2: 0
Aligner           : STAR
STAR Index        : s3://ngi-igenomes/igenomes//Homo_sapiens/Ensembl/GRCh37/Sequence/STARIndex/
GTF Annotation    : s3://ngi-igenomes/igenomes//Homo_sapiens/Ensembl/GRCh37/Annotation/Genes/genes.gtf
BED Annotation    : s3://ngi-igenomes/igenomes//Homo_sapiens/Ensembl/GRCh37/Annotation/Genes/genes.bed
Save prefs        : Ref Genome: No / Trimmed FastQ: No / Alignment intermediates: No
Max Resources     : 128 GB memory, 16 cpus, 10d time per job
Container         : singularity - nfcore-rnaseq.img
Output dir        : ./results
Launch dir        : /work/taubin/pipetest
Working dir       : /work/taubin/pipetest/work
Script dir        : /home/taubin/.nextflow/assets/nf-core/rnaseq
User              : taubin
Config Profile    : singularity

[warm up] executor > local
executor >  local (4)
[44/70808c] process > get_software_versions [100%] 1 of 1, failed: 1
[c0/d04379] process > trim_galore           [100%] 1 of 1, failed: 1
[0f/668817] process > fastqc                [100%] 1 of 1, failed: 1 ✘
[83/943e47] process > output_documentation  [100%] 1 of 1, failed: 1
[0;35m[nf-core/rnaseq] Pipeline completed with errors
WARN: Killing pending tasks (3)
ERROR ~ Error executing process > 'fastqc (SRR2532390)'

Caused by:
  Missing output file(s) `*_fastqc.{zip,html}` expected by process `fastqc (SRR2532390)`

Command executed:

  fastqc -q SRR2532390_1.fastq SRR2532390_2.fastq

Command exit status:
  0

Command output:
  (empty)

Command error:
  Skipping 'SRR2532390_1.fastq' which didn't exist, or couldn't be read
  Skipping 'SRR2532390_2.fastq' which didn't exist, or couldn't be read

Im' not familiar (at all) with singularity, the input files exist, the output files don't, if I launch myself the .command.sh it works, so it seems a problem of access.
May it be possible to have some help in usage?

Thanks in advance,

Aubin

[apeltzer]

That does sound a bit like your singularity installation doesn't have bind paths / automounts or similar configured. Can you share some details on where your files that you're trying to analyze are stored? Filesystem, mounted on a different (network?) storage?

For reference, it might be that Singularity can't read files from there, which you could also check by going into the container you downloaded and trying to cd to your workdirectory, e.g.

singularity shell nf-core-rnaseq.simg 
cd /path/to/your/work/dir

if you get an error there, it will be in 99% of all cases that your paths are not mounted properly to your Singularity container, thus causing it to "not see" your data. X-Ref info document on this: https://sylabs.io/guides/3.0/user-guide/bind_paths_and_mounts.html

[apeltzer]

As this stuff comes up relatively often, we might need to document this in our FAQ section @nf-core/core on the https://nf-co.re webpage :-)

[aubinthomas]

Thanks a lot for the quick answers!

I did the singularity shell. The Fastq files are in the current directory and can be read. If I go to the work directory I see the tmp subdirectories, and I can read the content of all the bash files. The only thing I can't read are the links to the input files.
For the system I'm using a public HPC, I'm trying to run it on gateway for testing, the storage is a shared storage (but I don't know more)

I will have a look of the documentation you sent me.
Thanks again!

aubin

[apeltzer]

The only thing I can't read are the links to the input files.

Are these links to a different filesystem? If yes - that filesystem should be in the singularity.conf as listed in the document I send you as a bind path and your issue should be gone 👍

[aubinthomas]

I did it on another server (simple with direct storage) where I can have a look on system, with singularity 2.5 installed.
The issue is the same on it.

When I add the directory to $SINGULARITY_BINDPATH with export, I see now the links with singularity shell.

Always with this env variable, when I launched nextflow, I have the same issue.
I imagine I have to enter this bind path elsewhere?

Sorry...

[drpatelh]

@aubinthomas Can you try and create a file called nextflow.config in the directory you are running the pipeline with the following content:

singularity {
  enabled = true
  autoMounts = true
}

Worth a shot...

https://www.nextflow.io/docs/latest/singularity.html#images

[aubinthomas]

Thanks @drpatelh ,
I created the file but nothing changed.

Here the last trace:

[warm up] executor > local
executor >  local (29)
[5f/54ceb4] process > trim_galore           [100%] 15 of 15, failed: 15
[86/c0ca53] process > fastqc                [100%] 13 of 13, failed: 13
[84/c7d16a] process > get_software_versions [100%] 1 of 1, failed: 1
[0;35m[nf-core/rnaseq] Pipeline completed with errors
WARN: Killing pending tasks (25)
ERROR ~ Error executing process > 'fastqc (SRR2532368)'

Caused by:
  Missing output file(s) `*_fastqc.{zip,html}` expected by process `fastqc (SRR2532368)`

Command executed:

  fastqc -q SRR2532368_1.fastq SRR2532368_2.fastq

Command exit status:
  0

Command output:
  (empty)

Command error:
  Skipping 'SRR2532368_1.fastq' which didn't exist, or couldn't be read
  Skipping 'SRR2532368_2.fastq' which didn't exist, or couldn't be read

Work dir:
  /poolzfs/ritchie/aubin/bolzniac/work/b3/bb27a371a2f438a818ff3c3eeac6db

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


/poolzfs/ritchie/aubin/bolzniac$ singularity shell nfcore-rnaseq.img
Singularity nfcore-rnaseq.img:/poolzfs/ritchie/aubin/bolzniac> ls
SRR2532315_1.fastq  SRR2532324_2.fastq  SRR2532334_1.fastq  SRR2532343_2.fastq  SRR2532353_1.fastq  SRR2532362_2.fastq ...

Singularity nfcore-rnaseq.img:/poolzfs/ritchie/aubin/bolzniac> cd work/b3/bb27a371a2f438a818ff3c3eeac6db/
Singularity nfcore-rnaseq.img:/poolzfs/ritchie/aubin/bolzniac/work/b3/bb27a371a2f438a818ff3c3eeac6db> ls
SRR2532368_1.fastq  SRR2532368_2.fastq

Singularity nfcore-rnaseq.img:/poolzfs/ritchie/aubin/bolzniac/work/b3/bb27a371a2f438a818ff3c3eeac6db> sh .command.sh

the last command works fine

[drpatelh]

Oh, you have it working now then?

[aubinthomas]

Nope, the problem is still the same.

[drpatelh]

Maybe this instead:

singularity {
  enabled = true
  autoMounts = true
  runOptions = "--bind /PATH/TO/BIND/PATH/" 
}

[aubinthomas]

OK it works, by changing directly the bind path inside singularity.conf
I'm happy!

Thanks @drpatelh and @apeltzer for your help, I learned a lot of things today .

[ewels]

x-ref https://github.com/nf-core/nf-co.re/issues/94