Using in silico click chemistry to design novel dopamine D3 receptor ligands as blood-brain barrier permeants

This repository contains supplementary data files and code for reproduction of the analysis in the paper. Clone this repository and follow the instructions to perform the library preparation and docking.

The file supplemntary_figures.pdf contains supplementary figures for the paper.

The analysis mainly requires the following software and packages:

• RDKit v. 2020.09.1.0
• Open Babel v. 3.1.0
• AutoDock Vina v. 1.1.2
• iDock v. 2.2.3
• other required python packages can be view in ipynb notebooks

Note: it is advised to create a separate python3 environment to run the prject scripts.

The data folder contains followng files:

• azides.txt, the library of azide molecules as SMILES strings
• click_compunds_dataset.csv, the dataset of click reaction products constructed with click_library_preparation.ipynb
• vina_data.csv, precomputed results of docking with AutoDock Vina
• idock_data.csv, precomputed results of docking with iDock

The workflow:

1. Library preparation. In order to prepare library by running in silico click reaction follow the steps in click_library_preparation.ipynb. It will produce click_compunds_dataset.csv with molecules ids, SMILES, logBB values computed by Clark and Rishton equations.
2. D3R docking

2.1 The docking is performed inside the docking dir. Note that this dir already contains the d3r.pdbqt receptor file prepared for docking as a rigid receptor with AutoDockTools4.

cd docking

2.2 First run docking for reference molecules (eticloprode and the scaffold). To do so, execute run_docking_for_reference.py. The sccript will run Vina and iDock on the reference molecules and produce a folder reference with epq and scaffold subdirs, containing the input pdbqt molecule files and output docking positions.

It is important to specify paths to your Vina and iDock executables inside the script (path_to_vina and path_to_idock variables). You can review specified Vina and iDock options inside the script. Edit --cpu, if you wish to change the number of cpus per job (default 20). Note that --exhaustiveness is set high to ensure thorough search, however it considerably increases the run time. --num_modes is set to 1 to pick the best binding position only. Note that the output parsing inside the script accounts for only one output mode. This notes are applicable for run_vina.py and run_idock.py too.

By default the docking is performed for 3 conformations of each molecule generated by Open Babel. This can be tuned inside the script in n_conf variable. The docking affinities for each conformation will be printed in terminal log during the execution:

python run_docking_for_reference.py

2.3 In order to prepare pdbqt 1,2,3-triazole ligands for docking execute run_obabel.py. The script will subset necessary compounds from data/click_compunds_dataset.csv and create the molecules dir with mol2 and pdbqt subdirs containing prepared conformations for 3062 compunds with logBB Rishton > 0.3:

python run_obabel.py

2.4 Vina run is initiated by run_vina.py. Don't forget to specify path to your Vina executable inside the script. You can review the Vina input options inside the script. The script will produce the vina_data.csv (precomputed available in data folder) with following columns: id (molecule id), isomer (either 14 or 15), conf_id (1 to 3 by default), affinity (kcal/mol). Docking positions will be stored in docking_vina_results dir, containing pdbqt and pdb subdirs with respective output file formats:

python run_vina.py

2.5 iDock run is initiated by run_idock.py. Don't forget to specify path to your iDock executable inside the script. It works quite same as run_vina.py. The script yields the idock_data.csv (precomputed available in data folder) with following columns: id (molecule id), isomer (either 14 or 15), conf_id (1 to 3 by default), idock_score (aka affinity in kcal/mol) and rf_score (pKd). Docking positions will be stored in new docking_idock_results dir:

python run_idock.py

4. Analysis of docking scores. Follow the analyze_docking_results.ipynb in order to aggregate and plot the docking output from vina_data.csv and idock_data.csv