[HOME]

NtileCodePredictor (ver. 0.4) - Predict the D² code sequences of given amino acid sequences (local structure prediction). It also predicts the amino acid sequences of given D² code sequences (protein design).
(About ver. 0.4 : Support different alternative location indicators for different chains in the PDB files.)
(About ver. 0.3 : It outputs not only a ".pred" file but also a ".code" file.)

2008-04-20 : fragment-to-5-tile-code conversion table of ASTRAL SCOP 1.73 95% is available [sample]
2009-02-28 : "frag_code1_full173.tbl" (fragments of length 1) is available.
2009-02-28 : "frag_code5_full173_ID.tbl (fragments with SCOP ID) is available. [sample]
2009-02-28 : "frag_code5_HIV1PR_ID.tbl" (fragments of HIV-1 protease variants) is available. [sample]
2009-02-28 : "frag_code5_full_HIV1PR_ID.tbl" (fragments of HIV-1 protease variants) is available. [sample]
    * 322 crystal structures and 120 NMR models of HIV-1 protease variants are used to compile the "HIV1PR" tables. [PDB ID list]

NtileCodePredictor [-f][-5][-7][-9][-1][-2][-3] [-v] [-d][-c] [-t table_name] [-h]  filename

If you type "NtileCodePredictor filename", you would obtain the predction based
on the "frag_code5_full.tbl" table.

Examples:

• Prediction based on "frag_code5.tbl" and creation of the ".udcode" file:
          % NtileCodePredictor -5 -v filename
  
  
• Prediction based on "frag_code5.tbl" and creation of the ".udcode" file:
          % NtileCodePredictor -5 -v filename
  
  
• Prediction based on "frag_code5_full.tbl" and "frag_code5.tbl" and creation of the ".udcode" file:
          % NtileCodePredictor -f -5 -v filename
  
  
• Protein design based on "frag_code15_full.tbl" and creation of the "_dsgn.udcode" file:
          % NtileCodePredictor -1 -v -d filename
  
  
• New structure check based on "frag_code21_full.tbl" and creation of the "_chck.udcode" file:
          % NtileCodePredictor -2 -v -c filename
  
  
• Prediction based on "frag_code5_HIV1PR_ID.tbl" (fragments of HIV-1 protease variants):
          % NtileCodePredictor -t frag_code5_HIV1PR_ID.tbl filename
  

(1a) A file of amino acid sequences (for local structure prediction):
            ".code" file created by the "ProteinEncoder" program (example),
             or    PDB ".ent" or ".pdb" file (example),
             or    FASTA type amino acid sequence ".seq" file (example).

(1b) A file of 5-tile code sequences (for protein design):
            ".code" file created by the "ProteinEncoder" program (example),
             or    ".code" file without "[Res.]" entries (example).

(2) Fragment-to-5-tile-code conversion tables:

[NOTE] "frag_code5_full.tbl" is recommended for local structure prediction.
               "frag_code15_full.tbl" and "frag_code21_full.tbl" are
                 recommended for protein design.

[NOTE] Conversion tables should be placed in the current directory.

[NOTE] The tables are the list of all fragments of the ASTRAL SCOP
               1.71-95 and their 5-tile codes:
                   - "frag_codeN.tbl" gives the 5-tile code determined by the
                      fragment of length N,
                   - "frag_code5_full.tbl" also gives the 5-tile code of the two
                      amino acids at both ends.

[NOTE] If the "-v" option is specifed, the corresponding PDB file should
               be placed in the directory where the input protein file exists.

(1) A ".pred" file
    It contains the predicted 5-tile code sequences and corrresponding
    statistics. The rate of successful prediction is also computed if the 
    input file is a ".code" file which conatins both "[Res.]" and "[Code]"
    entries (example).

    - A "xxx_dsgn.pred" file is created if the "-d" option is specified (example),
    - A "xxx_chck.pred" file is created if the "-c" option is specified (example),
    - A "xxx.pred" file is created otherwise (example1, example2).

(2) A ".code" file (example, example2)

(3) A ".udcode" file
    A ".udcode" file is also created if the "-v" option is specified so that
       one can view the result using the "ProteinViewer" program.

    - A "xxx_dsgn.udcode" file is created if the "-d" option is specified (example),
    - A "xxx_chck.udcode" file is created if the "-c" option is specified (example),
    - A "xxx.udcode" file is created otherwise (example1, example2).


[NOTE] Use "ProteinViewer" version 0.3 or later to view the ".udcode" files.

[NOTE] About "xxx_comp.udcode" (when viewed with "ProteinViwer")
         - Amino acids of failed prediction ("F") are denoted by large red spheres,
         - Amino acids whose code is contained in the prediction ("c") by large yellow spheres,
         - Amino acids without prediction by small blue spheres.
        (Example: pdb1rkl_comp)

[NOTE] About "xxx_accu.udcode" (when viewed with "ProteinViewer")
         - Amino acids with prediction accuracy less than 0.25 are denoted by large red spheres,
         - Amino acids with prediction accuracy less than 0.5 by large yellow spheres,
         - Amino acids without prediction by small blue spheres.
        (Example: pdb1rkl_accu)

[NOTE] About "xxx_comp_yyyy.udcode" (when viewed with "ProteinViewer")
         - Amino acids of successful predction ("-") are denoted by large blue spheres,
         - Amino acids whose code is contained in the prediction ("c") by large yellow spheres,
       
         - Amino acids with no hit (the corresponding code fragments are not
           included in the conversion table) are denoted by small red spheres (i.e., "new structure"),
         - Amino acids with hit count less than 5 by small orange spheres (i.e., "rare structure"),
         - Amino acids with hit count more than 999 by small blue spheres (i.e., "popular structure").
        (Example1: pdb1rkl_comp_dsgn, Example2: pdb1rkl_comp_chck)

[NOTE] About "xxx_accu_yyyy.udcode" (when viewed with "ProteinViewer")
         - Amino acids with no hit (the corresponding code fragments are not
           included in the conversion table) are denoted by small red spheres (i.e., "new structure"),
         - Amino acids with hit count less than 5 by small orange spheres (i.e., "rare structure"),
         - Amino acids with hit count more than 999 by small blue spheres (i.e., "popular structure").
        (Example1: pdb1rkl_accu_dsgn, Example2: pdb1rkl_accu_chck)

NtileCodePredictor reads a file of amino acid sequences and predicts the
5-tile code of the sequences (local structure prediction). It also predicts the
amino acid sequences of given 5-tile code sequences (protein design).

The following options are available: