I am using a local server, and when I use RNA STAR it does not allow me to select the reference genome, it appears as follows in the picture…
1 Like
I think above your screenshot you have to select to use a genome from your history since your local instance likely does not have pre-built indeces configured.
1 Like
The thing is that I want to use genome reference without built-in gene model but providing or not a .gtf; in both cases, you need to select a reference genome, and this is where the options are not displayed.
Once you select this, then if you place your .gtf file in the Gene model file for the splice junctions option, there are three different parameters.
Hi @Lilith_GM
I understand that there are a few ways to select a reference genome for that tool, but @marten 's main point is still likely correct: your server does not have reference genomes installed and indexed for this tool.
Inputs:
- reference genome (fasta from the history, or a server index)
- reference annotation (GTF,GFF3 from the history)
- Both combined into a single index (no reference data from the history, both available as a server index, together)
Using a custom reference genome (fasta) from the history is a choice. For a small genome like a bacteria, this can be fine. For a large genome like human, this is less preferred. A custom genome is indexed for the target tool every time the tool is run, then discarded. This can make jobs both run longer and consume more resources.
How-to is covered in many posts at this forum. The first hit with this search is a good reference.
You can get indexed reference genomes added to your server
→ directly (locally created one time, and used for mapping tools)
→ or through CVMFS (remotely created, and usually a mirror of what you find at public server, and these are linked into your server accessed locally for mapping tools).
Tutorials with the how-to
- Reference Data with Data Managers
- Reference Data with CVMFS without Ansible
- Reference Data with CVMFS
Then, the reference annotation is chosen to “match” the reference genome assembly. The details matter, and many posts are also at this forum about the topic.
(Technically, you could index the annotation along with the genome for this specific tool. Just be aware that that decision will mean creating your own index for this tool, and not taking advantage of the CVMFS pre-processed indexes, but you can still use those indexes for other tools.)
Hope that helps! Please ask if anything is not clear.