Genome incompatibility with STAR

Hello, I’m trying to use STAR to align my reads. I have successfully done this with the Zunla version of the pepper genome. I’m trying to repeat this with a different genome (CM334) that has better annotation and downstream resources. However, I get a tool error with this genome - ‘remote job server indicated a problem running or monitoring this job’. It is a custom genome but it has been correctly formatted and the annotation and genome chromosome names match. I get this error whether I run it with an annotation file or not so the problem seems to lie with the genome. Looking at the first few lines it looks ok but I don’t know about the rest of it. Is there a way to check the integrity of the genome or could there be any other reason for the incompatibility with STAR?

1 Like

Hi @alfrhul

It sounds like you have done good troubleshooting. Reference for others about those checks:

The problem could be with the genome formatting issue still present, or the tool itself be running out of resources. RNA Star is a memory-intensive tool. I’m guessing the second but we can confirm. Format or content issues are the usual cause for tools ending with a server-side/resource type of error.

There was another reported problem with this tool at around the same time frame that you wrote in. Their problem was the job not running at all. RNAStar not running

Meanwhile, for your case, perhaps start up a HISAT2 job using that same custom reference genome and annotation. Most mappers will not fail due to fasta format/content issues (when used alone) but incorporating reference annotation can reveal problems.

Run that job from the same history that the RNA Star failure is in, and please send in a bug report from one of the red error datasets. Please place a link to this Galaxy Help topic in the bug comments so we can link the two.

We are doing some checks administratively and will review bug reports around this tool. Feedback soon.

Issues with RNA-Star are resolved.

There were a few problems server-side, all now resolved. Primary ticket here:

If you are still having problems, there is an input problem or the job really is too large to execute at the public server.