The SRS-FastA subset similarity search at IUBio archive
  has been updated.
  
    - TFastA (aa query against translated DNA library) as well as 
      FastA searches are now available.
    - Protein libraries PIR, SwissProt and SwissNew are now searchable.
    - Source code patches are now available.
   
  This SRS-Fasta service is found at
    http://iubio.bio.indiana.edu:81/srsfasta/
    
  The source for creating your own SRS-FastA service is at
    <ftp://iubio.bio.indiana.edu/molbio/search/subsets/>


I have updated the SRS-FastA service and made source code patches
to SRS and FastA available.  It is not yet in a state where this will
be simple to install.  It will still require someone w/ basic
programming skills to put the patches in place and get the service
running.  I would like to make it easier to do, but I am running
low on the time I can put on this service in the short term.

I wish I could make this easier to have installed at your and other
sites. If you would like to install this, my best advice is
 
a) first install a standard, current SRS program with any nucleic,
protein databanks, if you haven't.  See EMBL at Heidelberg for this
(ftp://ftp.embl-heidelberg.de/pub/software/unix/srs/).  Make sure
it is working.  Also fetch the current FastA source, from
ftp://ftp.virginia.edu/pub/fasta/.  Note that I am working on sources
that are not the same as the current on.  I think my patches will
install okay on these, but they may not.

b) you may not want to install this SRS patch into a production
SRS server.  Though I have it on mine, and I am fairly sure it
can be added w/o messing up a current service, it would be safer to
create a separate SRS folder for this.  Install a second, fresh
SRS source distribution.  Follow instructions from the SRS-FastA notes
provided thru here (ftp://iubio.bio.indiana.edu/molbio/search/subsets/),
and apply the source code changes.  As I've noted, I work on version
4_02, whereas 4_08 or so is current. 

c) when this second SRS is installed and patched and tested for normal
operation, then test the SRS-FastA function and configure the Perl scripts
for your server.  

In this second SRS folder, just link your data/ and index/ folders 
from your production SRS folder.  To be safe, remove the srs/bin/osname/
files that can create new indices: remove everything but the
'getz' and maybe 'wgetz', and probably the 'odd' programs (to redo
the 'srssection' if you need to edit odd/ files here).  The SRS-FastA
program only uses this new 'getz' routine, but you will need to have
modified the odd/ files and rerun srssection according to notes. 

d) create a folder, e.g. /some/where/srsfasta, for the other
new files.  These could also be in your second srs/ folder.
I have a srsfasta/bin/ for fasta and tfasta, srsfasta/pub/ for
the .html forms, and srsfasta/source/ for the perl scripts plus
source code.   Your HTTP server needs to find the .html forms, and
your HTTP cgi-bin/ folder needs a link to the 'sfsubsearch' program
(see .html forms).  The Perl scripts need to find the srsfasta/bin/
fasta & tfasta, and to find the second srs/ folder tree with the
modified 'getz' program.

I test this by first by using the modified 'getz' to create a subset
library, as with

  getz  -l 'genbank gbnew' '[sq-org:drosophila*]'  -fse drosophila-na
            ^^ or embl or pir or other sequence library

This should create a subset library of all drosophila entries, in
the srs/tmp/ folder.  Create srs/tmp/ if you haven't, and set the
SRSTEMP environ var to point to it.  Or call it srs/subsets/ if you like,
the patches look for a folder set by environment "SRSTEMP" or by
parameter '-tdir folder-name'. If that works the SRS/getz part is okay.


Then I test the modified FastA & TFastA directly from command line,
giving them
   sequence:  some-test.seq
   library:  /path/to/srs/tmp/drosophila-na   12
                                              ^ type # for subset lib

Then I test using 'sfsubsection.pl' to create subset libraries
using the modified getz.   Finally I test with the HTTP server
to see if the forms and the 'sfsubsearch' is working fully.

For production be sure to have 'sfsubsection.pl' auto-run after your
SRS indices are updated, and have it update the 'sfsubsection.html'
file w/ new data counts.


-- Don Gilbert
software@bio.indiana.edu