Docking Automation Pipeline

Introduction

This project automates the process of docking ligands to proteins using various docking methods and scoring techniques provided by GNINA. It reads input data from a CSV file, converts necessary files, performs docking, and filters the results based on specified criteria, as explained below.

Usage

Upload the input CSV file (pairs.csv) containing protein, ligand, and site information to the csv/ directory. The CSV file should have the following columns:

• protein: Path to the protein PDB file.
• ligand: Path to the ligand PDB file.
• site: Path to the site PDB file.

Upload the required protein, ligand, and site PDB files to the proteins/ , ligands/ , and sites/ directories, respectively.

To run the docking automation script, use the following command:

./main.sh

Working

1. singularity exec containers/openbabel-3.1.0.sif ./docking_process.sh convert csv/pairs.csv

• CSV Parsing and File Conversion: The script reads the protein, ligand, and site information from the specified CSV file. It then converts PDB files to SDF format using Open Babel.

2. singularity exec containers/gnina-1.0.sif ./docking_process.sh gnina <docking_method> csv/pairs.csv

• Docking: Performs docking using the specified CNN scoring method using GNINA. - none: No CNN scoring is used. (fastest, traditional) - rescore: Rescoring using CNN. (fastest with CNN) - refinement: Refinement using CNN. (moderate speed. 10x slower than rescore) - metrorescore: Metropolis Monte Carlo sampling followed by rescoring using CNN. - metrorefine: Metropolis Monte Carlo sampling followed by refinement using CNN. - all: Ensemble of all CNN scoring methods are used. (slowest, extremely computationally intensive)

3. singularity exec containers/python-3.12.sif ./docking_process.sh unzip

• Unzipping: Unzips the docked results and stores them in the same directory.

4. singularity exec containers/python-3.12.sif python docking_process.py --mode filter --filter_type <filter_type>

• Filtering: Filters the docked results based on the specified filter type. - all: Retains all docked poses. - best: Retains the best docked pose based on the CNN score.

5. singularity exec containers/openbabel-3.1.0.sif ./docking_process.sh sdftopdb

• SDF to PDB Conversion: Converts the filtered SDF files to PDB format using Open Babel.

6. singularity exec containers/python-3.12.sif python docking_process.py --mode combine --filter_type <filter_type> --csv_file csv/pairs.csv

• Combining Ligand and Protein: Combines the best ligand conformation and given protein PDB files to generate a single PDB file.

7. singularity exec containers/plip-1.0.sif ./docking_process.sh plip

• PLIP Processing: Processes the combined PDB file using PLIP to identify interactions between the protein and ligand.

8. singularity exec containers/python-3.12.sif python docking_process.py --mode report

• XML Processing: Extracts the interactions from the PLIP XML file and generates a CSV file containing the interactions.

Dependencies

• Python 3.11
• RDKit
• Singularity
• Open Babel
• GNINA
• PLIP

SETUP

Singularity installation

Run the shell script located in the link below https://github.com/Mys7erio/scientiflow-singularity/blob/main/install-singularity.sh

Python and OpenBabel containers installation using Singularity

singularity pull containers/python-3.12.sif library://scientiflow/bioinformatics/python:3.12

singularity pull containers/openbabel-3.1.0.sif library://scientiflow/bioinformatics/openbabel:3.1.0

singularity pull containers/gnina-1.0.sif docker://gnina/gnina:latest

singularity pull plip.sif library://scientiflow/bioinformatics/plip:1.0

License

This project is licensed under the MIT License.