Tim Dudgeon wrote:
> I am wanting to set up SRS5 to index multiple blast files (pregenerated
> output, not generating the blast output within SRS).
>> I've made what I beleive should be the required changes by including a
> MULTI_BLAST database, based on what is already in blast.i and on how
> hssp.i handles multiple input files, but I'm getting run time errors:
>> MULTI_BLAST_DB:$library:[MULTI_BLAST group:@APPLICATION_LIBS
> comment:""
> format:@MULTI_BLAST_FORMAT maxNameLen:40 ifiles:{"blast.i" "blast.is"}
> type:user
> # files:{
> # $file:tmp
> # }
> ]
the problem is that at the moment there is the choice between 1) each
entry
is in a separate file and 2) entries are contained in one or more
flatfiles.
in the latter case the file names as attribute "files" of $library MUST
be specified. Searching files with a search expression as for hssp
(*.hssp)
work only for the single entry/file case. That means to fix the problem
quickly you need to remove the above comments and write something like
files:{
$file:tmp1
$file:tmp2
...
}
...obviously that is not what you want since the names and number fo the
blast output files will probably vary quite a lot. To make that a little
more flexible you could write a script that generates the 'files' part,
inserts
it into blast.i and runs srssection ....sorry about that!
Since I am working on the problem with the multiple directories for pdb
i will probably fix the multiple blast file problem next since the too
are quite related (at least as seen from SRS's perspective).
regards
Thure
problems