Inge Jonassen, Dept. of Informatics, University of Bergen, Norway
You receive a tar file containing the executable program(s) and this documentation.
In order to use the program, you need to unpack the tar file by writing for example
$ tar xvf Spratt22.tar
The following files will be produced:
The first two are executable programs, the third is this file.
You also need the dssp program dsspcmbi that can be obtained separately.
SPratt2 is run from the command line. The format of the command line is as follows:
spratt2 <filename> <maxdist> <minangle> <min_support> <max_rmsd>
where filename is the name of a file containing PDB file names
maxdist is in angstrom max dist between
minangle is the minimum cosine value
min_support is the minimum number of structures to match a pattern
max_rmsd is the maximum rms deviation between a pattern and a match
Additional parameters can be given.
- List of structures:
- SPratt2 takes as input a file containing a list of filenames/PDB identities. It is
assumed that the structures to be analysed are present in the working directory each
with a filename identical to its PDB id. PDB ids are by convention four characters
long. If only one chain from a PDB entry is to be read, the identity of the chain
should be given as a fifth letter, e.g., 1flmA identifies the file 1flm, chain A.
An example of an input file to SPratt2:
SPratt2 also reads a dssp output file for each structure. In the example, SPratt2
will look for 1hud.dssp, 1gcb.dssp, 1fie.dssp, 1iam.dssp, 1ppn.dssp.
This can be produced by dsspcmbi that must be obtained separately and run prior to SPratt2.
In order to avoid the time consuming parsing of PDB and dssp files each time the program is
run, SPratt2 will produce binary files containing the necassary information about each
structure. In the example, the first time SPratt2 is run, it will read the pdb and dssp files
and produce binary files 1hud.bin, 1gcb.bin, 1fieA.bin, ... The next time the same structures
are analysed, the binary files will be read.
- This is the maximum distance (in angstroms) between two residues for
them to be considered as neighbours.
Typical values are 8, 10, 12.
- To reduce the lengths of neighbour strings and still keep the residues most
likely to be interacting, i.e., residues facing eachother. Residues i and j are
facing each other if angle i and angle j both have cosine values larger than or
equal to minangle.
Angle i is the angle between the vector from c alpha i to
c alpha j and the vector from c alpha i to mean side chain atom i.
Angle j is the angle between the vector from c alpha j to
c alpha i and the vector from c alpha j to mean side chain atom j.
If minangle=-1, all residues within maxdist from residue i will be included in
residue i's neighbour string. If minangle=0, all residues within distance maxdist
that are facing roughly towards i (and i towards them) will be included in residue
i's neighbour string.
- This is the minimum number of structures to match a structure pattern.
- This is the maximum root mean square deviation to be accepted when superposing
a possible pattern match onto the pattern.
The postp program takes as input an output file from SPratt2 and makes a summary
postp is run from the command line. The format of the command line is as follows:
postp <pattern file> <pdbids file> <scop file> [options]
where pattern file is a spratt2 output file
pdbids file is a file with the IDs of the structures analysed by spratt2
scop file is a file with scop IDs for the pdbids
subset options: -nr - patterns ranked by support
-sc - patterns ranked by score
-cover - gives subset of patterns covering structure set
-all - include all patterns
output format: -html - produce html output
-latex - produce latex output (not yet supported)
First running SPratt2:
spratt2 t 10 -1 3 1 > t.out
This produces output file t.out, then run the post-processing program:
postp t.out t dir.dom.scop.txt_1.50 -cover -sc -html > t.html
br>which produces the output file t.html which can be viewed using
any web browser, e.g., Netscape or Internet Explorer.