Automating ADFR Suite & Meeko for Molecular Docking With Windows Batch Scripts
Citation: Lee, S. (2024, March 14). Automating ADFR Suite & Meeko for Molecular Docking With Windows Batch Scripts. Sepo.
March 14th, 2024
Introduction
Molecular docking is a technique with great potential in the field of medicine. This is exemplified by its power to discover the drugs clofazimine and benidipine for Chagas disease [1]. It allows researchers to predict the preferred orientation of a ligand as it binds to a target receptor, how powerful or likely the bind is, and identify the most reactive parts of the receptor [2].
AutoDock for Flexible Receptors (ADFR) is a revised version of Autodock Vina and Autodock 4 with similar capabilities as its predecessors [3–7]. But has better performance in calculating the docking of flexible receptors [3–7].
Meeko is a Python package developed to prepare small molecules for molecular docking with Autodock [8].
However, processing multiple ligands against a single receptor can be time-consuming without the right automation strategies. This article discusses using Windows batch scripting to automate receptor preparation, ligand preparation, grid box creation, and actual docking.
Method
Preparing Your Environment
Before automating your docking workflow, ensure that the ADFR suite and Meeko are installed and correctly configured on your system. This includes having all necessary executable files from ADFR Suite ( prepare_receptor, agfr, and adfr) and Meeko (mk_prepare_ligand.py) accessible in your system's PATH, to invoke these tools from any directory within the Command Prompt.
Script Overview
The script performs several key steps in sequence:
Receptor Preparation: Initially, the script prepares the receptor molecule, a step that includes adding hydrogens and other necessary preprocessing tasks to ensure the receptor is ready for docking.
Positive Control Ligand Preparation & Target File Generation: The script prepares the known positive control ligand and then generates a target file with information about how it will react with the receptor.
Ligand Processing Loop: For each ligand file in the specified directory, the script:
Converts the ligand from
.sdfto.pdbqtformat, making it compatible with the ADFR suite's docking protocols.Executes the docking simulation between each ligand and the receptor usign the target file generated by the positive control ligand, providing insights into the potential binding modes and affinities of the ligand-receptor complex.
Script
@echo off
setlocal enabledelayedexpansion:: Prepare the receptor for docking
prepare_receptor -r receptor.pdb -A checkhydrogens -o receptor.pdbqt:: Prepare the positive control ligand and generate the docking grid
:: Assume the positive control ligand file is named "positive_control.sdf"
echo Processing positive control ligand: positive_control.sdf
python mk_prepare_ligand.py -i "positive_control.sdf" -o "positive_control.pdbqt"
agfr -r receptor.pdbqt -l "positive_control.pdbqt" -o grid
echo Finished preparing grid using positive control:: Loop through each .sdf ligand file in the directory, excluding the positive control
for %%f in (*.sdf) do (
if not "%%f"=="positive_control.sdf" (
echo Processing ligand: %%f
:: Convert ligand from SDF to PDBQT
python mk_prepare_ligand.py -i "%%f" -o "%%~nf.pdbqt"
:: Execute the docking simulation using the pre-generated grid
adfr -t grid.trg -l "%%~nf.pdbqt" -J "docking_%%~nf"
echo Finished processing ligand: %%f
)
)
echo All ligands have been successfully processed.Running the Automation Script
To run the automation:
Save the script with a
.batextension. An example would beautomate_docking.bat.Place this batch file in the directory containing your ligand
.sdffiles and thereceptor.pdbfile.Open the Command Prompt, navigate to the directory containing your script, and execute it by typing its name.
Understanding Results
The only essential outputs that need to be understood is the target file (.trg ) and docking output.
To view grid.trg , type about grid.trg after you have navigated to the directory it is located in through the Command Prompt.
It should look similar to this:
docking target file
date : JOHN Jan 14 18:18:24 2024
node : JOHNB880
AGFR : v1.2rc1
AutoSite : v1.0
receptor : receptor.pdbqt
FlexRec : None
covBond : No
box : mode ['ligand'], padding 4.00
center : 0.372 -0.834 -0.306
length : 15.000 14.250 10.500
size : 0040 0038 0028
spacing : 0.375
maps :
types : A Br C Ca Cl d e F Fe G GA H HD HS I J Mg Mn N NA NS
OA OS P Q S SA Z Zn W
W map : weight 0.60 entropy -0.20
gradients: Yes, kept largest negative cluster
pocketMode : ['ligand']
#fillpts : 174 points
file : C:\Users\John\docking\grid\translationPoints.npyThe important part of this file is the box information. It gives information about the coordinates of the active site. This information can be used for docking with other programs like Autodock Vina. The other components are not as important to actual molecular docking and more importantly to molecular dynamics and assessing the accuracy of the program.
For the output from molecular docking, the mode that is of interest is usually just mode 1. In mode 1, the affinity (kcal/mol) score is what is relevant because it shows how likely the ligand is to bind to the receptor. The more negative the affinity value is, the stronger the bond. This is the best value to rank and compare the ligands for their docking results [8].
Validation of Results & Limitations
It should also be noted that this protocol has its limitations. The protocol does not account for factors like water molecules that can affect actual binding. Therefore, validating the results with other measures is important.
It also does not account for certain ligands like macrocycles that require special settings for molecular docking. To dock them, consult the literature and manuals for the software used in this article.
To validate the results of docking, if possible, it is best to use a ligand/s that have already been shown to have reactivity with the receptor to use as a control group [9].
Actual experimental validation is also another option. There also appear to be other ways to validate results, but I am unfamiliar with them. Therefore, it is best to consult the literature if you need other methods to validate results.
Conclusion
In summary, molecular docking, particularly with enhancements in AutoDock for Flexible Receptors (ADFR) and Meeko, offers significant promise for drug discovery, as demonstrated by its application in identifying treatments for Chagas disease [1]. The use of Windows batch scripting for automation streamlines the process, allowing for efficient processing of multiple ligands against receptors. While this computational approach accelerates the discovery phase, it’s crucial to remember that these predictions require experimental validation to confirm their real-world efficacy. This methodology not only showcases the power of computational biology in advancing medical research but also highlights the necessity of integrating computational predictions with empirical evidence to drive forward the development of new therapeutics.
References
Bellera, C. L., Balcazar, D. E., Vanrell, M. C., Casassa, A. F., Palestro, P. H., Gavernet, L., … & Talevi, A. (2015). Computer-guided drug repurposing: identification of trypanocidal activity of clofazimine, benidipine and saquinavir. European journal of medicinal chemistry, 93, 338–348.
Forli, S., Huey, R., Pique, M. E., Sanner, M. F., Goodsell, D. S., & Olson, A. J. (2016). Computational protein-ligand docking and virtual drug screening with the AutoDock suite. Nature protocols, 11(5), 905–919. https://doi.org/10.1038/nprot.2016.051
Zhang, Y., Forli, S., Omelchenko, A., & Sanner, M. F. (2019). AutoGridFR: Improvements on AutoDock Affinity Maps and Associated Software Tools. Journal of Computational Chemistry, 40(32), 2882–2886.
Zhang, Y., & Sanner, M. F. (2019). AutoDock CrankPep: combining folding and docking to predict protein–peptide complexes. Bioinformatics, 35(24), 5121–5127.
Ravindranath, P. A., & Sanner, M. F. (2016). AutoSite: an automated approach for pseudo-ligands prediction — from ligand-binding sites identification to predicting key ligand atoms. Bioinformatics, 32(20), 3142–3149.
Ravindranath, P. A., Forli, S., Goodsell, D. S., Olson, A. J., & Sanner, M. F. (2015). AutoDockFR: advances in protein-ligand docking with explicitly specified binding site flexibility. PLoS computational biology, 11(12), e1004586.
Zhao, Y., Stoffler, D., & Sanner, M. (2006). Hierarchical and multi-resolution representation of protein flexibility. Bioinformatics, 22(22), 2768–2774.
meeko. (2023, July 28). PyPI. https://pypi.org/project/meeko/
Cosconati, S., Forli, S., Perryman, A. L., Harris, R., Goodsell, D. S., & Olson, A. J. (2010). Virtual Screening with AutoDock: Theory and Practice. Expert opinion on drug discovery, 5(6), 597–607. https://doi.org/10.1517/17460441.2010.484460