Running on with SGE

[ningb]

Hi,

Thank you for developing this pipeline!

This is not an actual issue, but more of a question. Our institution uses a SGE cluster computing system, with CentOS. Could you please let me know what might be the easiest way to modify and run run your pipeline on such system?

Thanks so much!!

Best Regards

[drpatelh]

Hi @ningb . Youre welcome 😃

I havent personally used SGE but I guess thats the beauty of Nextflow and the way these pipelines are written! You should just be able to create a separate custom.config file that has the details for your preferred execution. The docs for the Nextflow SGE scope are here:
https://www.nextflow.io/docs/latest/executor.html#sge

For example, I have created one for my Institute because we use Slurm.
https://github.com/nf-core/configs/blob/master/conf/crick.config

You can then just add -c custom.config to the command-line and it should work.

Let me know how you get on.

[ningb]

Hi @drpatelh Thanks for your fast response!

I added in the SGE related options to the config file and tried to run it as you suggested. However, one issue I encountered is that I can't run the --skip_trimming command (my reads were cleaned and trimmed by the sequencing facility). Nextlow would give me error

ERROR ~ No signature of method: java.util.ArrayList.ifEmpty() is applicable for argument types: (ArrayList) values: [[]] Possible solutions: isEmpty(), isEmpty(), first(), first() -- Check script 'main.nf' at line: 1469 or see '.nextflow.log' file for more details

Have you ever experienced this before?

Thanks again!

Best,
Boting

[drpatelh]

Hi @ningb . Thats actually a bug in the pipeline 😅 Sorry.

It will be fixed in the next release. These two lines:

chipseq/main.nf

Lines 499 to 500 in 21be314

	ch_trimgalore_results_mqc = []
	ch_trimgalore_fastqc_reports_mqc = []

Should be changed to:

    ch_trimgalore_results_mqc = Channel.empty()
    ch_trimgalore_fastqc_reports_mqc = Channel.empty()

You should be able to edit those lines in your local version of the pipeline ~/.nextflow/assets/nf-core/chipseq/main.nf.

Ive submitted a PR for the fix too #121.

Thanks!

[drpatelh]

@ningb This will now be fixed in the next release or you can run the dev version of the pipeline by adding -r dev to the pipeline command. You may have to delete your local version of the repo first though because it has since been updated.

rm -r ~/.nextflow/assets/nf-core/chipseq/

Closing for now but let me know if you have any further problems 👍