Cloud Computing for Assembly?

[peromhc]

Is anyone working on this? I am interested in using cloud computing for a de novo Solexa assembly of mammal transcriptome. Looked at Amazon, but they will only dedicate ~15gb RAM per instance/user.

Any ideas? Matt

[rglover]

Would it be possible to start up multiple instances and cluster them so they share the RAM? I could be talking rubbish though, as I'm not well versed in EC2 or computing clusters!

There's a test program in the Amazon Elastic Map Reduce called CloudBurst which is supposed to do short read mapping to a reference genome, although I haven't used it, I've only seen the paper. The Amazon info for it is at http://developer.amazonwebservices.c...xternalID=2272.

[ECO]

That is pretty cool. I'd love to get together a more complete userguide for launching EC2 instances for beginners (myself included).

I know of another similar cloud project that is getting ready to launch...maybe that time is getting closer!

[spenthil]

I am starting to look into this, I'll let you guys know if I get anywhere.

[kkamerath]

Hi Matt,
I've talked to Amazon Web Services and it sounds like a viable option for running a distributed-RAM method like ABySS. Pretty affordable too. They recommended a contractor to me who has experience setting up MPI-enabled jobs on EC2 which would be necessary to do ABySS.

I believe they are also going to get some larger-RAM machines available in the near future.

Best,
Kristen

[ramcob]

Just go and try it here erp on cloud

[qtrinh]

Hi,
You can use the "High-Memory Quadruple Extra Large Instance" instance type, which comes with 64GB of RAM and costs $1.8 per hour:

http://aws.amazon.com/ec2/instance-types/

For prices, see http://aws.amazon.com/ec2/pricing/

[qtrinh]

Hi,
If there are enough interests on a user guide on how to get starting with Amazon EC2, I can put one together.

Q