Native, Amazon EC2 of n3phele

[Darwin Sorensen]

We are relative newcomers to Qiime but have successfully used Qiime VirtualBox to work through the tutorial and use alpha diversity tools on a couple of 454 16S reads files.  I have merged triplicate sample 454 files to create files with up to 65k reads.  We would like to do beta diversity analysis on these community samples but QVB runs out of memory in our installation.  What do you recommend for a Qiime system big enough to do this: native installation on a workstation with an Intel Core i5-2400, quad core processor, 8 GB RAM, and 500 GB hard drive; Amazon EC2 image or n3phele?

Thanks.

Darwin

[Jose Carlos Clemente]

Darwin,

for your current needs, a native install should suffice. If you get
into larger datasets (e.g. millions of reads for 454, or HiSeq), you
might want to give it a try to EC2.

Jose

[Greg Caporaso]

Just my two cents on this: I think using the Amazon EC2 images is a
great way to go as it saves you from the trouble of doing a native
QIIME install. I generally recommend the EC2 images as the starting
point for QIIME users, and I use these for the majority of my
analyses.

Greg