Extracting and formattiing by script

Steps for collecting and formating a SAM_T99 server prediction:

1) Get the server results for the target sequence as a text file from the
   CAFASP2 meta-server at

    http://cafasp.bioinfo.pl/target/

   Using the results stored on the CAFASP meta-server will help
   ensure that the results we submit are the same as the results
   on the meta-server. Save the file in the dir from.cafasp. The naming
   convention used so far is

       from.cafasp/cafasp.samt99.t????.txt

   where the ???? is the target number.

2) Make a directory to place the formatted results in.
   Run the script prep-to-submit.perl to extract the database hits
   and secondary structure prediction from the meta-server text file.
   The script should be run as

    prep-to-submit.perl \
        -runname <runname> \
        -fromserver <server file from.cafasp/cafasp.samt99.t????.txt> \
        -method <method documentation filename> \
        -registration <casp registration number filename> \
        -scopdata <scop summary file> \
        -checktarget <good target sequence file>
        -targetoffset <target_aa_index_offset>\   #default value of 0 
        -filtertheomodels [on|off] \              #default value of on
        -filterscop [on}off]       \              #default value of on
        -maxmodels max_models      \              #default value of 5

   The -fromserver argument should be the raw server output file from
   the meta-server obtained in step 1.

   The -method argument documents the methods used to generate the
   prediction. The standard file is in methods.fold.

   The casp4 registration code for SAM_T99 is 2000-8506-6614,
   and is in the file registration.fold in this directory.

   The -scopdata argument should be a file containing the scop classifications
   of all pdb structures. The file we will be using is at

       /projects/compbiodata/scop/dir.dom.scop.txt_1.50

   The -checktarget argument should be a file containing the target
   sequence that is known to be good. This file is used to check that
   the target sequence has not been corrupted in any of the pairwise
   alignments. For casp4, these known good files should be in

       /projects/compbio/experiments/casp4/t????/T????.seq

   where the ???? are the target number.

   The -targetoffset can be used to adjust the indexes if the target indexes
   should start at something other than 1 (for example T0100 has its first
   24 residues truncated, so this argument should be 24 for T0100).

   The -filtertheomodels is for enabling or disabling filtering out hits
   that are theoretical models. Similarly, -filterscop is for filtering
   out redundant scop superfamilies. By default, both are set to "on";

   The -maxmodels argument is for setting a limit on how many models should
   be submitted. The default is 5.

   An example of the scripts use is:
    
    ../prep-to-submit.perl \
        -runname t0090 \
        -fromserver ../from.cafasp/cafasp.samt99.t0090.txt \
        -method ../methods.fold \
        -registration ../registration.fold \
        -scopdata /projects/compbiodata/scop/dir.dom.scop.txt_1.50 \
        -checktarget /projects/compbio/experiments/casp4/t90/T0090.seq

   The output of stderr and stdout is the table of database hits extracted
   from the meta-server raw output. Listed with the pdb id of each hit
   it its evalue, scop classification, and whether the hit is to be submitted
   or rejected, along with the reason why.

   Among the outputs of the script will be the file <runname>.submit
   which contains the file to email to the Prediction Center. 

3. Check the summary and any warning message output from the script.
   If you use the theoretical model filter, you may get warning messages
   for hits whose pdb files have no EXPDTA record. These may need to be
   checked by hand to see if they contain real experimental models.
   The summary will list all the hits and whether they should be submitted
   or rejected along with the reason why.

   Check that the files <runname>.<hit id>.a2m.pa
   and <runname>.<hit id>.al2fasta.a2m have essentially the same alignment.

4. Mail the <runname>.submit.al files and <runname>.submit.ss file
   to submit@predictioncenter.llnl.gov


Manual Formatting

If you need to run the procedure by hand because of something the
prep-to_submit.perl script cannot handle, follow the CASP3 README for
submission procedures.

Some minor changes to the procedure involve additional command line
arguments to some of the formatting programs.
The argument -checktarget should be given to the fasta2al and al2fasta
programs if target sequence checking is needed. This argument should be
the filename of the target sequence in fasta format. For CASP3, it would
 automatically look in the pce/casp3 subdirectories, but it is now a
command line argument.

The format-prediction script now accepts three additional command line
arguments, which are:

       -registration <registration code file>
       -model <model number>                this is optional, by default 1
       -targetoffset <target offset>        optional, by default 0