Dockit

Perform high-throughput molecular docking with multiple targets and ligands using AutoDock Vina series engines.

💎 Features

• Molecular docking with multiple target and ligands at the same time.
• Flexible residue declaration for targets.
• Use other Vina-like engines e.g. Qvina2/Qvina-W or Smina.
• Original PDBQT preparation method as distributed with MGLTools.
• Automated PDBQT file preparation.
• Optional ligand energy minimization using obminimize.
• Output CSV file with all docking results for all modes for an easy access to the docking results.
• Docker support. Run it anywhere!

🚀 Installation (Docker or Conda)

Docker

Dockit can be run with Docker. You must have Docker & docker-compose installed.

git clone https://github.com/aretasg/dockit
cd dockit

Conda

Install conda first

git clone https://github.com/aretasg/dockit
cd dockit
conda env create -f environment.yml
conda activate dockit

💻 Example usage

1. Determine the search box size and centre positioning using Chimera (Tools > Structure/Binding Analysis > Vina) or similar.
2. Define search box and docking parameters for each target in dockit_param.csv.
3. Copy protein and ligand PDB files into targets/PDB and ligands/PDB folders, respectively and run:

python app/dockit.py

or if using Docker

docker-compose up

• The calculation will take some time depending on the parameters and the number of files. Results can be found in the results folder.
• Visualise the results with a molecular viewer of your choice by loading ligand PDBQT file in the results folder and target PDB or PDBQT file in targets directory.
• CSV file dockit_results.csv is generated in results folder with all the docking results.
• Run with -r flag to reset to the pre-run state - PDBQT and result files will be removed.

Argument description in dockit_param.csv

Argument	Description	Required
target	file name of the target located in targets/PDB - excluding the extension	❗
x_center	X coordinate of the seach box center	❗
y_center	Y coordinate of the search box center	❗
z_center	Z coordinate of the search box center	❗
x_size	search box size in the X dimension (Angstroms)	❗
y_size	search box size in the Y dimension (Angstroms)	❗
z_size	search box size in the Z dimension (Angstroms)	❗
engine	type of engine to use to run the docking (e.g. vina or qvina2)	❗
exhaustiveness	exhaustiveness of the global search (roughly proportional to time, default=8)	🤔
num_modes	maximum number of binding modes to generate (default=9)	🤔
seed	explicit random seed	🤔
cpu	number of CPUs available to use for docking (default is to auto detect the number of cores available)	🤔
energy_range	maximum energy difference between the best binding mode and the worst one displayed (default = 3 kcal/mol)	🤔
flex_resi	specify target residues to be treaded as flexible during docking	🤔

FAQ

• Dockit will dock every target against every ligand in targets/PDB and ligands/PDB, respectively.
• Dockit is not an input for docking preparation tool. Input files have to be in PDB format. Conversion to PDBQT adds hydrogens and removes non-polar ones.
• Edit dockit_param.csv to change any parameters to be run with the docking engine, including search box parameters.
• Run with -m flag perform energy minimization for ligands using obminimize with default settings.
• Supports declaration of flexible residues for targets. Please look at dockit_param.csv for an example for how to declare flexible residues.
• Supports other Vina-like engines e.g. Qvina2. Please specify the engine name or path to it in the engine column of dockit_param.csv. Dockit comes installed with vina, qvina2, qvinaw and smina. You should be able to use any other engine with the same CLI as Vina.
• Specify custom scoring system weights in the weight_* fields of dockit_param.csv. More about the scoring system here.
• Dockit utilises threading to distribute multiple simulations across available cores. Please change the cpu argument to a finite value to ensure distributed processing.
• The necessity of selecting the charges (Kollman/Gasteirger) is obsolete with Vina due to the scoring system being based on hydrophobic and hydrogen bond interactions compared to its predecessor AutoDock 4. Nevertheless, the default charge for targets and ligands are set as Kollman and Gasteirger, respectively.

✏️ Authors

Written by Aretas Gaspariunas. Have a question? You can always ask and I can always ignore.

🍎 Citing

If you found Dockit useful for your work please acknowledge it by citing this repository.

License

MIT License.

Acknowledgments & Disclaimer

prepare_* files are distributed as part of MGLTools 1.5.6 and all the ownership is credited to their respective authors (Morris et al., 2009).