Low computation usage on AWS using Galaxy

[adhalleran]

Hi Everyone,

I am attempting to analyze HiSeq 2500 RNA-Seq data using a Galaxy instance run through AWS, but the analysis is taking an unexpectedly long time (to run around 10M reads through Tophat using XenTro2 as the reference genome has taken over 3 hours and no end is in sight.)

Looking at the core usage on AWS monitor it appears that the Galaxy instance is only using 8% of a c3.8xlarge (with autoscaling set to allow it to use up to four c3.8xlarge if needed).

Is this normal? Is there a way to push Tophat to use more of the computing resources available through AWS?

Thanks

[GenoMax]

Are you using a prebuilt galaxy image? It is may be configured to use only one core for tophat. If you are not comfortable using command line/have good knowledge of aws it may be more trouble than it is worth to go modifying configs.

You may want to install tophat directly in your aws instance and use that instead of galaxy.