There is two significant features of the PSI/PHI Blast server. First, it requires that the FASTA sequences from which the BLAST database is formatted to contain gi numbers in their sequence id fields and be formatted with '-o T'. Second, it allows interactive iteration of searches based on recalculated position specific score matrix (PSSM). Most of the FASTA protein sequences from NCBI FTP site have gi in their defline. We recommend that one use the preformatted protein databases from NCBI whenever possible.
Local databases may not be used with this version of PSI/PHI Blast unless they have ">gi|12345..." prefix in the definition lines. One possible solution is to modified the deflines to convert them to NCBI format with "pseudo" gis using custom script. Databases for the PSI/PHI Blast should always be created with formatdb using the '-o T' flag. Test database 'test_aa_db' was created with this flag while the database 'test_na_db' was created without.
If this distribution is installed not in the "/blast" directory under HTTPD documents root directory, then path to the distribution should be set by an environment variable WWW_ROOT_PATH in the file psiblast.cgi or psiblast_cs.cgi