README.md

Description

Finding similarities and redundancy in chemical data sets.

Installation

git clone https://github.com/moozeq/dd-redun.git

cd dd-redun
git submodule update --init --recursive
pip3 install -r requirements.txt

Prepare database

Ligands

1. Download PDBBind database (e.g. CASF-2016) and move its coreset to dd-redun/coreset (you may also used pre-built database from demo/db.smi, in that case skip to 4.)

   # move coreset from CASF-2016
   mv CASF-2016/coreset .
   
   # or use pre-built database from demo and skip to 4
   cp demo/db.smi .

2. In case of CASF-2016 you may need to remove 4mme complex, because ligand from this complex is causing error when creating fingerprints:

   rm -rf coreset/4mme

3. Run script from below (simply getting smiles and id for each ligand):

   for f in `ls coreset/`; do obabel -imol2 coreset/${f}/${f}_ligand.mol2 -osmi | awk '{print $1" "$2}' >> db.smi; done

4. Database should be at dd-redun/db.smi
5. For docking functionality whole coreset folder must be under dd-redun/coreset directory

Receptors

6. Run script from below (simply merging all x_pocket.pdb files into one file database) or use pre-built database from demo/prots.pdb:

   # generate prots database
   for f in `ls coreset/`; do cat coreset/${f}/${f}_pocket.pdb >> prots.pdb; done
   
   # or use pre-built database from demo
   cp demo/prots.pdb .

7. Database should be at dd-redun/prots.pdb
8. Build G-LoSA using clang or g++:

   g++ glosa.cpp -o glosa

Requirements

Main functionality

• openbabel

Scaffolds

• strip-it

Docking

• AutoDock Vina
• ODDT
• PDBBind coreset

Receptors

• G-LoSA

Usage

Help

./redun.py -h
./scorun.py -h
./sredun.py -h

Database file

./redun.py db.smi
./scorun.py db.smi [ints]
./sredun.py prots.pdb

In pipeline

cat db.smi | ./redun.py
cat db.smi | ./scorun.py [ints]
cat prots.pdb | ./sredun.py

Report

Report from project is available here (in Polish).

Selected results plots