Chapter 10: Virtual Screening with AutoDock Vina & Smina
Chapter 10: Virtual Screening with AutoDock Vina & Smina
Virtual screening computationally docks libraries of thousands to millions of small molecules against a target receptor to prioritize the top candidate compounds for experimental in vitro synthesis and binding assays.
Vina estimates the free energy of binding \(\Delta G\) (in kcal/mol) using an empirical scoring function balancing steric interactions, hydrophobic contacts, and hydrogen bonding:
\Delta G_{\text{bind}} \approx \sum_{\text{pairs}} \left( f_{\text{gauss}} + f_{\text{repulsion}} + f_{\text{hbond}} + f_{\text{hydrophobic}} \right) - N_{\text{rot}} \Delta G_{\text{torsion}}
More negative \(\Delta G\) values indicate tighter thermodynamic binding affinity. A score of < -8.0 kcal/mol typically indicates sub-micromolar binding potency.
1. Running AutoDock Vina
# Execute AutoDock Vina on 16 CPU cores
vina --config config.txt --out docked_poses.pdbqt --cpu 16
# Inspect output log:
# mode | affinity | dist from best mode
# | (kcal/mol) | rmsd l.b.| rmsd u.b.
# -----+------------+----------+----------
# 1 | -9.4 | 0.000 | 0.000
# 2 | -8.8 | 1.421 | 2.104
# 3 | -8.5 | 2.150 | 3.2802. High-Throughput Screening with Smina
Smina is a high-performance fork of AutoDock Vina supporting custom scoring functions and batch SDF input files.
# Screen 10,000 compounds from an SDF library against a receptor
smina \\
-r receptor.pdbqt \\
-l drug_library.sdf \\
--autobox_ligand crystal_ligand.pdb \\
--autobox_add 6 \\
--exhaustiveness 8 \\
--cpu 32 \\
-o screened_hits.sdf.gz