Molecular Docking with Explicit Waters¶

Using AutoDock Vina 1.2.x, Meeko, and AutoGrid4.

Objective: perform molecular docking while explicitly considering water molecules in the binding site.

Recommended method: AutoDock4 scoring with --scoring ad4 and explicit waters attached to the ligand workflow.

1. Prerequisites and Installation¶

1.1 Install AutoDock Vina 1.2.x¶

# 1. Install compiler and dependencies
# Ubuntu/Debian
sudo apt update
sudo apt install build-essential libboost-all-dev swig

# 2. Download and compile Vina
git clone https://github.com/ccsb-scripps/AutoDock-Vina.git
cd AutoDock-Vina/build/linux/release
make

# 3. Put the executable in PATH
sudo cp vina /usr/local/bin/
# or add it to PATH:
# export PATH=$PATH:/path/to/AutoDock-Vina/build/linux/release

1.2 Required Scripts¶

Download the Meeko, molscrub, and AutoDock helper scripts:

https://github.com/forlilab/Meeko/release/meeko/cli/mk_prepare_receptor.py
https://github.com/forlilab/Meeko/release/meeko/cli/mk_prepare_ligand.py
https://github.com/forlilab/molscrub/blob/develop/scripts/scrub.py
https://github.com/ccsb-scripps/AutoDock-Vina/develop/example/autodock_scripts/mapwater.py
https://github.com/ccsb-scripps/AutoDock-Vina/blob/develop/example/autodock_scripts/dry.py

1.3 Other Required Programs¶

• PDB2PQR for pH/protonation preparation
• PyMOL for inspection and conversion
• AutoGrid4, usually available with MGLTools or AutoDockTools

2. Receptor Preparation¶

# 1. Fix and protonate the receptor at the desired pH, here pH 5.0
pdb2pqr --ff=AMBER --ph-calc-method=propka --with-ph=5.0 --keep-chain --pdb-output receptor_pH5.0.pdb receptor_original.pdb receptor_pqr.pdb

# 2. Prepare the receptor for AutoDock with Meeko
python3 mk_prepare_receptor.py -i receptor_pH5.0.pdb -o receptor --box_center 12.0 9.0 -1.0 --box_size 25.5 31.5 25.5 -p -g

Generated files:

• receptor.pdbqt
• receptor.gpf
• receptor.box.pdb, useful for visual inspection

Open receptor.box.pdb in PyMOL to confirm that the docking box is well positioned.

3. Ligand Preparation¶

In PyMOL:

load ligand.pdb
save ligand.sdf

Then run:

scrub.py ligand.sdf -o ligandH.sdf

Prepare the hydrated ligand:

python3 mk_prepare_ligand.py -i ligandH.sdf -o ligand_hydrated.pdbqt -w

4. Generate Affinity Maps with AutoGrid4¶

# Check ligand atom types
grep "^ATOM\|^HETATM" ligand_hydrated.pdbqt | cut -c 78-79 | sort | uniq

# Edit receptor.gpf and keep only the atom types that exist in the ligand.
# Common example for organic molecules:
# ligand_types C OA HD H

# Run AutoGrid4
autogrid4 -p receptor.gpf -l receptor.glg

5. Create the Special Water Map¶

python3 mapwater.py -r receptor.pdbqt -s receptor.W.map

6. Run Docking¶

For hydrated docking, use AutoDock4 scoring:

vina --ligand ligand_hydrated.pdbqt --maps receptor --scoring ad4 --exhaustiveness 32 --out ligand_out_ad4.pdbqt

7. Post-process and Classify Waters¶

python3 dry.py -r receptor.pdbqt -m receptor.W.map -i ligand_out_ad4.pdbqt

The output file has the suffix _DRY_SCORED.pdbqt and classifies waters as:

• STRONG: well-retained water
• WEAK: marginal water
• DISPLC: likely displaced water

8. Final Visualization¶

pymol receptor_pH5.0.pdb receptor.box.pdb ligand_out_ad4_DRY_SCORED.pdbqt

Important Files¶

Step	Main generated file	Purpose
Protonation	receptor_pH5.0.pdb, receptor_pqr.pdb	Corrected and protonated receptor
Receptor preparation	receptor.pdbqt, receptor.gpf	AutoDock receptor files
Ligand preparation	ligand_hydrated.pdbqt	Ligand with explicit waters
Maps	receptor.*.map, receptor.W.map	Affinity maps and water map
Docking	ligand_out_ad4.pdbqt	Docking results
Post-processing	ligand_out_ad4_DRY_SCORED.pdbqt	Final result with classified waters

Final Notes¶

• Always use --scoring ad4 for this hydrated docking workflow.
• Increase --exhaustiveness to 32-64 for fragments or flexible ligands.
• This method is not recommended for virtual screening of thousands of compounds.