Elastic BLAST vs. BLAST cloud

Question

[Joseph Kashmer]

Hello, I am a student trying to create a project program which requires executing BLAST search functions. I am currently using the module from Biopython (NCBIWWW) which executes a url request to the BLAST cloud server. However, it is painfully slow and NCBI BLAST AMI is deprecated as per their site.

I have been considering using BLAST+, but doing so seems to be impractical since it requires heavy computational resources, given that my program function is only simple. So, I have been digging on Elastic BLAST, which apparently performs searches on the cloud servers as well, and is much faster.

However, I am not familiar with this. Can I run it as a subprocess in my program?

Answer 1

[GenoMax]

You can use locally installed blast+ to launch remote blast searches on the command line (what you are thinking of as blast cloud but in reality is Web Blast). This is not meant to be a replacement for large amount of blast+ searches that should be run locally. If you tried to use very large sequences (number and size) then you will likely end up getting errors.

Elastic BLAST on the other hand is meant to be run completely in the cloud on VM's. There is no local component. It is supported on AWS and GCP. While you may be able to connect to your instance of elastic BLAST remotely, you would be required to set elastic BLAST up in the cloud and pay for the cloud resources. It would not be free.

[Yannick Wurm]

Another alternative (if you don't want to create AWS or GCP accounts but are still prepared to pay some cash) is through the SequenceServer API. Example code here.