Maker Annotation Pipeline AWS AMI

Question

[njue]

Anyone know where I can find an AWS AMI for Maker for genome annotation? Be great to not have to build one myself.

Answer 1

[Charles Warden]

I didn't think installation was that bad for MAKER. However, I was only annotating a small sequence - if you are working with something larger, I could see how you might want to have more computational resources and/or have better parallelization (and I think there might have been something about the threading option that I didn't exactly get working).

I had a bit of a learning curve on AWS (spending a few hundred dollars within a month: probably still less than a formal course, but worth considering). However, if you already have experience with knowing how to minimize compute time and other good stuff, that probably doesn't matter.

I think Google Cloud still offers a $300 credit to get started. I want to say that I encountered issues with requesting more than 24 cores in certain zones. However, if you can't run the job locally, and only need to use the cloud once (and it is your first time), maybe that is worth considering?

Nevertheless, if you can find a Docker image with MAKER, that may help in all of these situations (local, AWS, or Google Cloud). However, I think there was one caveat that I didn't exactly get worked out on Google Cloud (if using Docker+gcsfuse Bucket storage), even though I could run DeepVariant through another strategy (already set up as a pipeline for that particular program).