No options available (Select Reference Genome)

, , , , , ,
1

Hi everyone,

We recently installed the Galaxy Server to our mainframe. However, we are having difficulty in using RNA STAR as there is no reference genome available (see attached picture). How can we rectify this?

Screenshot 2022-06-13 at 17.10.20
Screenshot 2022-06-13 at 17.10.202052Ă—1616 302 KB

Many thanks,
Chris

2

Did you added references already?

If not maybe it is good to check out this first:

2 Likes
3

Is there any video tutorial on how we can do this? We just simply want to add the reference genome (human) with built-in index and without the built-in gene model.

Apologies, we are very novice to this.

Thanks,
Chris

4

Hi @cw951

The slides that @gbbio shared are part of a larger tutorial suite. Admin topics are here:

Review the section “Data Management & Reference Data”. In short, you have several choices for the method. These include:

  1. Create your own indexes using Data Managers. These are like other tools Galaxy, but only available for local server admins. You’d need to install the tools, then access them under the “Admin” tab, then either run the tools in a specific order directly or use our batch processing utility.
  2. Link in or copy of the indexes available at public Galaxy servers from our data server. This would be configured by a local admin. Human genome builds are included.

The how-to is covered in the tutorials and some include videos. I also added a few tags to your post that cover prior Q&A about this topic (troubleshooting, etc). If you get stuck or have problems, feel free to ask more questions.

Each genome build is associated with

  • the genome fasta and several technical indexes. Some optional, some required. Example: items like the Samtools index would be needed to interpret BAM outputs produced by mapping tools.
  • and tool specific indexes (including RNA-Star for most).

Having a built-in indexes is different from using a Custom genome/build where just the fasta is add to the working history. Why? Galaxy creates those technical and tool specific indexes at runtime as single-use temporary files. Custom genomes/builds are fine for smaller genomes (or any “-ome”) but larger genomes would require much more time/resources to create those indexes every time the tool is run. Convenient for some use cases but not practical for others.

For human, definitely use a built-in indexes. And if you chose to use the pre-computed indexes we host, you won’t need to scale up your local resources to create the indexes with Data Managers. You will still need to have sufficient resources to execute a tool. For `RNA-star, I think the minimum working memory would be about 60 GB – but you can check with your own data. Run the tool at a public server with a representative dataset query and against a genome build target, click into the Job Details ( :information_source: icon within the result dataset), and review the resources used in the lower part of the report.

Hope that helps!

1 Like