Sequence Libraries

OMSSA searches protein libraries formatted for BLAST.  A precompiled set of popular libraries such as nr can be found here (note that Swiss-Prot is implemented as a "mask" on nr, so you need to download both sets of files).  You can also use the program formatdb to create your own protein sequence libraries from various formats, such as fasta.  formatdb can be found in the BLAST executable distribution -- look for an archive that begins with "blast" and is specific to the operating system that you use. A sample command line is "formatdb -i sequences.fasta -o F -p T ". This will will create a set of files that contain the sequences and other information for speedy access.

The library files can be placed in the same folder as the OMSSACL executable, but ideally should be kept in a separate folder.  If you wish to keep the libraries in another folder, you need to tell OMSSA where to find the blast libraries. The simplest way is to specify the folder when specifying the library, e.g. "omssacl -d /path/to/your/blast/libraries/library_name" where library_name is the name of the library without the ".p*" extensions.

Another way to use another folder is to specify the follow the operating system-specific instructions below:

 1. On Linux and MacOS X you need to have a file called .ncbirc in your home folder which points to the blast library folder.  Example file contents:

[BLAST]
BLASTDB=/path/to/your/blast/libraries

Please do not put quotes around the path name. 

 2. For Windows you need to have a file called ncbi.ini in your WINDOWS or WINNT folder which points to the BLAST library folder.  Example file contents:

[BLAST]
BLASTDB=c:\\path\\to\\your\\blast\\libraries

Please do not put quotes around the path name. Note that backslashes are repeated.

Note that some libraries are implemented as a "mask" (aka subset) on another library.  You can tell if a library is a mask if there is a file that ends with .msk.  In this case, you will need the library that is being masked.  For example, in the precompiled set, swissprot is a mask on nr, which means you need to download both swissprot and nr to search swissprot.