cathoud / ppigremlin Goto Github PK

This document describes how run ppiGReMLIN. ppiGReMLIN is a scalable graph based strategy which models
protein-protein interfaces as graphs.

PREREQUISITES

ppiGReMLIN requires Python 3.4.3 or above along with a few Python packages, which can be found in the
file "requirements.txt".

DEFAULT FILES

atom_types.csv: contains the pyhsicochemical properties of atoms from each of the 20 amino acid residues
used in contact prospection.
interactions.csv: interatomic distance criteria used to compute interactions. Specifies the distance
range for each interaction as well as the atoms properties.

RUN PPIGREMLIN

1.	Create a file containing a list of the  PDB ID's of the structures to be used as input to ppiGReMLIN,
each line containing a single PDB ID along with the chains to be used in contact calculation.

Example: structure 1EJM (CRYSTAL STRUCTURE OF THE BPTI ALA16LEU MUTANT IN COMPLEX WITH BOVINE TRYPSIN)
contains 6 chains named from A to F. The chain pairs (A,B), (C,D) and (E,F), each represent a
protease-inhibitor complex, from which we want to prospect contacts in the interfaces. The line entry in
the file for this PDB strucuture would be as following:

1EJM,A,B,C,D,E,F

By default the program will search PDB for the structures at the input list and download them.
Alternatively, the user can place the structures in the "pdbs" directory.

Files "soybean.txt" and "bcl.txt" are the input files for the Serine protease and BCL-2 sample datasets
respectively.

2.	In Terminal, from the ppiGReMLIN directory, run:

	$ python main.py <pdbidsfile> <directory>

		<pdbidsfile>: the file contaning the PDB ids from the strucutures used as input
		<directory>: name of subdirectory in ppiGReMLIN to place the results

Data generated for ppiGReMLIN visualization tool will be placed in a directory called "data", inside
the results directory specified.