Using and running the nf-core rnaseq pipeline on a server or cluster

Question

[diego1530]

Hi, I know that the rna-seq pipeline from nf-core has been used a lot here and could you help me out https://nf-co.re/rnaseq/usage

In this case, I would like to know if when we install the pipeline on a server, is it necessary for us as users to manually install each program that requires the pipeline or we just download the pipeline on our server and automatically the programs are also installed depending on a container?

This is a serious concern, as there is no information about it. I am anxious to hear your comments.

Answer 1

[Phil Ewels]

Please see the "pipeline software" section of the documentation: https://nf-co.re/usage/installation#pipeline-software

(Short answer - no, the pipelines come with all software packaged for you already).

[diego1530]

Thanks Phil for your kind reply. If I want to run any pipeline online, what is the process, do you have any documentation?

[Phil Ewels]

I'm not really sure what you mean by running a pipeline online?

You can use the online GUI for entering pipeline parameters with the _Launch_ button on the pipeline page (direct link for nf-core/rnaseq is https://nf-co.re/launch?pipeline=rnaseq ). Once completed you can then launch the pipeline using Nextflow directly, nf-core/tools or Nextflow Tower.

[diego1530]

Dear Phil, I'm running the nf-core/pipeline offline in a server and I know I have to download the human genome using AWS-iGenomes in FASTA and GTF formats. My goal is to use the --aligner 'star_rsem' option and according to documentation, both the STAR and RSEM indices should be present in the path specified by --rsem_index (see #568). It means that both indices must have the same path. Now, for getting these indices, I have to use manually following command to build RSEM references: 'rsem-prepare-reference' or the pipeline itself could generate these indexes automatically with the '--save_reference' option? I'm confused on this part and hope listening to you.

[Phil Ewels]

The pipeline will generate the indices if you do not provide them, all it needs is fasta and gtf. If you want to prevent it from doing this every time then yes, use --save_reference and then move the generated indices somewhere logical where you can use them for the next run with --rsem_index.

Note also that you can just download the indices and move them to your cluster if you prefer: https://ewels.github.io/AWS-iGenomes/

Let's continue any further discussion on the nf-core slack, where the whole community can help out (faster and better answers than just me on Biostars). See https://nf-co.re/join There's a channel dedicated to the nf-core/rnaseq pipeline: #rnaseq.

[Phil Ewels]

Please also mark my answer as the solution to this post if you don't mind :)