Basic Local Alignment Search Tool (BLAST) is a popular sequence similarity search tool. The National Center for Biotechnology Information (NCBI) maintains a public BLAST server at blast.ncbi.nlm.nih.gov. The NCBI BLAST server is intended mainly for interactive use, with the expectation that users will perform a moderate number of searches (i.e., 10-20) per day. Some users have many more searches to perform than can be accommodated at the NCBI site. For such batch users, the NCBI offers the stand-alone BLAST+ package that can be run on local hardware. Running the stand-alone package requires that the user have access to sufficient compute resources (e.g., a local cluster) as well as some degree of familiarity with IT issues and command-line applications. Additionally, some users do not need to perform a large number of searches on a regular basis, but may need to do so occasionally. A cloud provider may offer an acceptable solution for these batch users.

The NCBI provides a BLAST+ server image at Amazon Web Services (AWS) and Google Compute Engine (GCE). The BLAST server image includes the BLAST+ applications, a "FUSE" client that can cache databases from the NCBI (during the first search), a PERL script that mimics a subset of the NCBI URL API, and a simplified BLAST search webpage. The instance runs an http daemon for the remote BLAST searches. The FUSE client can download a limited set of databases (see the FUSE client). NCBI does not not charge for use of the software, but any instances will incur the cloud provider's charges.

  • The AWS Amazon Machine Image (AMI) can be run directly via the AWS Marketplace.
  • The Google Compute Engine deployment process can be started from the GCE Cloud Launcher; documentation can be found here.

The BLAST server image supports three different search methods:

  1. Users can ssh into the instance and run stand-alone BLAST+ at the command-line.
  2. The instance has a simple webpage for search submission.
  3. The instance supports the NCBI-BLAST Common URL API interface.

The latter two methods are remote searches that return a "Request ID".

The remote search methods allow one user in a group to register at a cloud provider, configure the instance, and handle payment for the running instance(s). Other members of a group can then use the instance through the URL API or webpage without their own cloud provider account. The URL API interface is ideal for users familiar with the BLAST URL API at the NCBI. Currently, the BLAST server image runs searches on only one instance. Many simultaneous users, through the remote interface, could overload a single instance. The NCBI is working on an approach that distributes searches to multiple instances. Information and links to the latest BLAST AMI are available at the BLAST Searches at a Cloud Provider web page.