Best deployment method of NGS tools on a cluster?

[GenoMax]

SGE is the default queuing system on our rocks cluster. We run all jobs through SGE. Primary (CASAVA, PacBio SMRTanalysis) and secondary data analysis.

BioTeam has info on using SGE. This link should get you started: http://bioteam.net/2009/09/sge-training-slides/

awesome , so you are using SGE with Bowtie? what tools are using with it?

[rskr]

Switching to Hadoop seems interchangeable and I'm not sure about having a hadoop cluster if we decide we want run other applications on it. Any advice folks have is muchly appreciated. Thanks!!
-Rich

From what I have seen hadoop doesn't play well with other types of apps, and most informatics apps don't really need the hadoop functionality. Though at a super high level meeting of business managers and non-computational-biologists, you might be able to sell hadoop as a easy and simple way to do high throughput computing in a map reduce framework using general purpose hardware and cheap java labor

[greener]

awesome , so you are using SGE with Bowtie? what tools are using with it?

[GenoMax]

Rich: We have been using SGE with rocks for some time. But PBS should work as well.

I also wanted to put a plug in for "modules" project. It makes management of multiple software versions/environment a breeze. It works well with Rocks http://modules.sourceforge.net/

[greener]

Howdy everyone,
We are building a small rocks cluster (4-5 nodes to start) and I am hoping to configure it to perform some denovo tools (abyss and trinity) but also to be able to run the tuxedo suite(bowtie, tophay, cufflinks). The question I have is there any way to deploy bowtie runs across machines (using something like SGE or PBS)? It seems like using myrna or crossbow would require making the cluster a hadoop cluster correct? Switching to Hadoop seems interchangeable and I'm not sure about having a hadoop cluster if we decide we want run other applications on it. Any advice folks have is muchly appreciated. Thanks!!
-Rich