Configuring STAMP

To use STAMP, the user must set the environment variable STAMPDIR to the full path of the subdirectory `/defs' in which the installation was made. The directory containing the STAMP binaries, which is STAMPDIR/bin should also be included in the user's PATH environment variable.


STAMP reads PDB coordinate information and DSSP secondary structure assignments. Thus, you should have copies of the PDB and DSSP files for the structures in which you are interested (although DSSP files are not strictly required).

STAMP input files do not require that the full paths of the PDB / DSSP files being loaded should be specified. As an alternative to a full path, STAMP can find PDB and DSSP files for a domain by using only the domain identifier and sets of patterns defined in the files `STAMPDIR/pdb.directories' and `STAMPDIR/dssp.directories'. The format of each line in these files is:

    <directory>  <prefix>  <suffix> [RETURN]

For example, the default pdb.directories file looks like this:

    _ _ _
    ./ _ _
    ./ _ .pdb
    ./ _ .pdb.Z
    ./ _ .pdb.gz
    ./ pdb .ent

STAMP searches for the PDB/DSSP files corresponding to a domain by taking the first four characters of the domain identifier as a PDB code and combining the code with each of patterns in turn to construct a test file name. If the test file exists, then that is used as the source of PDB coordinates or DSSP records. The fields in the pattern on each line are:
1. Directory path
2. File prefix
3. File suffix
If a field has the value `_', then it is ignored when creating a test filename. For example, suppose STAMP is searching for the PDB file for a domain with the identifier 4chaa. Using the default pdb.directories file, STAMP will attempt to open the following sequence of files:
4cha ./4cha ./4cha.pdb.Z ./4cha.pdb.gz ./pdb4cha.ent
The first file which it finds will be loaded to find coordinates for the domain.
If you specify the full path to the PDB files in a STAMP input file, or the PDB files are in the directory in which you run STAMP, then the default pdb.directories file will be sufficient and you need not modify it.
A recent modification (version 4.2) is to look in each of the `distr' type sub-directories for filenames. Some people store PDB files in a format, e.g.

    <directory>/ab/pdb1abc.ent

Where the two letter sub-directory name corresponds to the second two characters in the four letter PDB code (i.e. ignoring the leading number). STAMP now handles these file types. If you just specify the top directory, the program will explore suitable two-letter sub-directories corresponding to each file it is looking for.

dssp.directories contains a description as to where possible DSSP files may be found. The format is as for pdb.directories, e.g.

    _ _ _
    ./  _ .dssp
    ./  _ .dssp.Z

For example, the DSSP file for 4mbn might be found in the file ./4mbn.dssp

STAMP now reads compressed files (.Z or .gz suffixes). In order for this to work properly, you must have the programs zcat (.Z) and gunzip (.gz) installed on your system.