7.2. Why are BLAST searches submitted from a CLC Workbench to NCBI taking so long?
Some parameter choices can affect how long a BLAST job will take, but in the case of launching searches that will run on the NCBI Servers, the time it takes for BLAST searches to run and complete will greatly depend on how busy the NCBI BLAST Servers are. When submitting BLAST searches to the NCBI BLAST Servers using CLC Workbenches, you will likely see better execution time during off-peak hours. (Unfortunately, it is hard to predict exactly when these hours are, but generally speaking, outside US office hours is probably better than within them.)
Our tests indicate that the time required for a search submitted from CLC Workbenches to the NCBI to be complete can vary substantially, and that these times do not necessarily reflect how long the same job would take if running directly at the NCBI BLAST website. The NCBI controls the availability of server time and prioritization of BLAST jobs submitted via different routes. These are not things we have control over.
If you are running a BLAST search using a single sequence list with many sequences, then one way you could try speeding things up would be to split up your sequence list into a few smaller sequence lists, and then submit each of those lists as a separate BLAST search job. It may be useful to note, when deciding if this is something you wish to do, that the results for each BLAST job you launch are separate and that there is currently no way to merge these outputs within the Workbench.
Alternatively, you may consider running your BLAST against a local database.
Relevant manual pages regarding BLAST and splitting of Sequence List are:
CLC Main Workbench:
CLC Genomics Workbench: