Installing Sun Unix Version of ProteinProspector

1. If you wish to access the programs over an Intranet you'll need to have a service such as TCP/IP set up with Domain Name Service (DNS) running. Network access also facilitates moving sequence databases onto your computer. In order to run the programs directly from a single computer, however, network access is not necessary.

2. If your machine is not already configured as a web server, create a directory (ex: /home/httpd) to locate the web server software. While you may wish to run your web server from the /usr filesystem, these directions suggest using the /home filesystem, which typically is allocated much more disk space. You can of course run your server from /usr and provide a symbolic link to the /seqdb directory located elsewhere.

   Make sure you have enough disk space for your databases.

   As of Jan 15th 1999 the uncompressed sizes of the public FASTA databases supported by ProteinProspector were:

   • dbEST 1033 MB
   • Genpept 136 MB
   • NCBInr 147.5 MB
   • Owl 110 MB
   • SwissProt 34 MB

   By Sep 27th 2000 the corresponding database sizes were as follows:

   • dbEST full, 837 MB compressed, estimated 2.79 GBytes uncompressed
   • dbEST human, 352 MB compressed, estimated 1.17 GBytes uncompressed
   • dbEST mouse, 215 MB compressed, estimated 717 MBytes uncompressed
   • dbEST others, 268 MB compressed, estimated 893 MBytes uncompressed
   • Genpept 98.8 MB compressed, 236 MB uncompressed
   • NCBInr 135MB compressed, 256 MB uncompressed
   • Owl 65.2 MB compressed, 124 MB uncompressed - last updated May 22nd 1999
   • SwissProt 21.9 MB compressed, 39 MB uncompressed
   • Ludwignr 228 MB all elements uncompressed

   The sizes of the individual elements of the Ludwignr database were as follows:

   • gp_nr.seq 20.0 MB
   • gpu_nr.seq 1.90 MB
   • pir1_nr.seq 959 KB
   • pir2_nr.seq 5.40 MB
   • pir3_nr.seq 6.66 KB
   • pir4_nr.seq 37.2 KB
   • swiss_nr.seq 41.7 MB
   • swiss_varsplic_nr.seq 1.41 MB
   • trembl_new_nr.seq 40.2 MB
   • trembl_nr.seq 115 MB
   • trembl_varsplic_nr.seq 378 KB
   • wormpep_nr.seq 245 KB
   • yeastpep_nr.seq 140 KB

   In addition to this the files made by the FA-Index program are between 8-16% of the size of the database.

3. If your machine is already configured as a web server, skip to the next step. If your machine is not already configured as a web server, download web server software to the /home/httpd directory and install according to the directions provided with the web server. This will probably be the most time consuming part of the installation. Note: in these instructions the document root is /home/httpd/htmldocs. Server software is available from:
   Freeware
   • NCSA httpd server ( mostly UNIX i.e. Sun, SGI etc.)
   • EMWAC https server ( Windows NT )
   Free only to academic and non-profits
   • Netscape servers (all platforms)
   You may wish to create a home page for your server, the ProteinProspector file mshome.htm can serve this function. If you create a symbolic link from a file called /home/httpd/htmldocs/index.html to home/httpd/htmldocs/mshome.htm, this page will then be located at the URL http://yourcomputer.yourinstitution.com.
   If you do NOT have root privileges on your workstation, you should configure your server to run from a port >1023, rather than the standard port 80. Port 8080 is a common configuration, and is set in the file /home/httpd/conf/httpd.conf. The URL of your server will then be: http://yourcomputer.yourinstitution.com:8080.
   From the /home/httpd directory, start your server with a command like httpd -d /home/httpd and verify that you can access your home page.

4. Place the Prospector distribution file (file name below) precompiled for your system to your document root directory /home/httpd/htmldocs.
   • SunOS 4.1.3 and higher: prospector__sun.tar.Z.

   Untar and uncompress the Prospector distribution set using the commands 'uncompress prospector__sun.tar.Z' and tar xvf 'prospector__sun.tar'.
   You should then find the following:

   • Readme.txt
   • mshome.htm
   • results_mspattern
   • results_msfit
   • results_msseq
   • results_mstag
   • • results_mspattern (symbolic link)
     • results_msfit (symbolic link)
     • results_msseq (symbolic link)
     • results_mstag (symbolic link)
   • • gifs
     • instruct
     • js
     • misc

5. Check to see that the files /home/httpd/htmldocs//*.cgi have execute permission. If not, change the permissions to make them executable: 'chmod +x *.cgi'.

6. Create the directory /home/httpd/htmldocs//seqdb, then install some databases and create the accompanying index files.

7. This completes the installation. TIME TO TEST DRIVE. Access one of the input forms either from the links on your homepage or at the URL: http://yourcomputer.yourinstitution.com:8080/msfit.html and press the submit search input button. If you haven't been able to install the databases, the MS-Product and MS-Comp programs will still work, enabling you to test that you've installed everything correctly. You must address the programs via an http:// type URL to provide the CGI syntax that invokes the executable.

   If your server generates an error message about access not being allowed to the programs, then you may need to do the following:

   Modify your server configuration to allow applications with a .cgi suffix to be run inside the server's document root directory. For the NCSA httpd and Apache servers this is accomplished by adding the following lines to the conf/srm.conf file:

   ScriptAlias // /home/httpd/htmldocs//
   AddType application/x-httpd-cgi .cgi