General tips for using BLAST on Abel

From mn/ibv/bioinfwiki
Revision as of 13:34, 23 February 2015 by Ralfne@uio.no (talk | contribs) (Created page with "= General tips for using BLAST on Abel  = <div>Most programs in the NCBI BLAST package will print out a help file if writing the program name followed by “-help”, for...")

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

General tips for using BLAST on Abel 

Most programs in the NCBI BLAST package will print out a help file if writing the program name followed by “-help”, for instance:

blastn –help

The program name followed by only “-h” will print out a summary of the help file. It may be a good idea to save the help file as a text file which then can be consulted at need:

blastn –help>blastn_help.txt

It is quite useful to be able to see this help file during the writing of a blastn command. You can do this by simply opening another terminal window, displaying the help file in one window, while writing the blastn command in the other window.

Likewise, it is a good idea to test out BLAST commands on a minimal data set before running a full-scale BLAST search. For instance, you can include only one or two sequences in your query file, instead of thousand of sequences. Such a test run will reveal if the correct modules have been loaded, if the input files are correct and if the output shows up where you expect it to. 

A simple BLAST search against the “nr” database takes around two minutes, as we have seen. It is possible to do such small searches directly from the Abel promt (i.e. without submitting a job script). However, longer tasks should not be started directly from the promt, as this would cause the system to hang. Therefore, tasks running longer than 30 minutes will automatically be killed.

If running longer BLAST searches, you have a number of options:

Submit a BLAST search job script

As this means that your search will enter the Abel queue system, you need to specify your account, the time the program needs to run, the amount of memory necessary etc. See the Abel help pages for details.

Use “qlogin” to directly interact with Abel during job execution

The creation of job scripts may seem quite complicated at times; the “qlogin” functionality allows you to enter the commands exactly as you have seen above, while still getting the computing resources necessary to run massive BLAST searches. The drawback is that you will have to wait in front of your computer till a sufficient slot of computing time is available. This often happens inside minutes for shorter jobs, but may become too prolonged for larger jobs. See the Abel help pages for further details.

Use the “freebee” system

This system gives you unrestricted access to your files on Abel, and supports all the programs also included on Abel, without computing time limits. This system, however, is not as fast as Abel; heavy calculations may take too much time to complete on “freebee”.