Molecular Docking Engine

Bioshift provides comprehensive molecular docking capabilities through AutoDock Vina integration, supporting both CPU and GPU-accelerated versions for drug discovery and virtual screening applications.

AutoDock Vina Capabilities

Comprehensive molecular docking with multiple Vina versions and GPU acceleration.

Multiple Vina versions (1.1.2, 1.2.3-1.2.7) included

GPU acceleration with Vina GPU 2.1 (CUDA)

Automated receptor and ligand preparation

Batch docking for virtual screening

Grid search optimization

Result analysis and visualization

Available Vina Versions

Vina 1.1.2

CPU

Original AutoDock Vina version

Features:
  • Classic scoring function
  • Well-validated
  • Widely cited
Best for: Standard docking, publications

Vina 1.2.3

CPU

Updated scoring function

Features:
  • Improved accuracy
  • Better hydrogen bonding
  • Updated parameters
Best for: General purpose docking

Vina 1.2.5

CPU

Latest stable CPU version

Features:
  • Optimized performance
  • Bug fixes
  • Enhanced stability
Best for: Production workflows

Vina GPU 2.1

GPU (CUDA)

GPU-accelerated for high-throughput

Features:
  • 10-50x speedup
  • Batch processing
  • CUDA optimization
Best for: Virtual screening, large libraries

Docking Workflow

Standard molecular docking workflow from receptor preparation to result analysis.

Step 1

Receptor Preparation

Convert protein structure to PDBQT format

Nodes:
File InputReceptor Preparation
Step 2

Ligand Preparation

Prepare ligands with charges and torsions

Nodes:
SDF ReaderLigand Preparation
Step 3

Grid Definition

Define search box around binding site

Nodes:
Grid Search
Step 4

Docking Execution

Run molecular docking simulation

Nodes:
AutoDock Vina / Vina GPU
Step 5

Result Analysis

Analyze binding poses and scores

Nodes:
Docking AnalysisProLIF Interaction
Step 6

Visualization

View protein-ligand complexes

Nodes:
Web 3D ViewerStructure Draw 2D

Grid Box Parameters

Key parameters for defining the docking search space.

ParameterDescriptionDefault Value
center_x, center_y, center_zGrid box center coordinates (Å)Calculated from ligand or pocket
size_x, size_y, size_zGrid box dimensions (Å)20 × 20 × 20
exhaustivenessSearch thoroughness8 (higher = more thorough)
num_modesNumber of binding poses9
energy_rangeEnergy range for poses (kcal/mol)3.0

CPU vs GPU Performance

Performance comparison for virtual screening of compound libraries.

Library SizeCPU Time (Vina 1.2.5)GPU Time (Vina GPU 2.1)Speedup
10 compounds5 min30 sec10x
100 compounds50 min2 min25x
1,000 compounds8 hours15 min32x
10,000 compounds3 days2 hours36x
100,000 compounds30 days20 hours36x

* Performance may vary based on hardware, protein size, and docking parameters.

Best Practices

Recommendations for optimal docking results.

Receptor Preparation

  • Remove water molecules (except structural waters)
  • Add missing hydrogens at pH 7.4
  • Optimize hydrogen bond network
  • Check for missing residues/atoms

Grid Box Setup

  • Center on known binding site or co-crystal ligand
  • Add 10Å padding around reference ligand
  • Use exhaustiveness ≥8 for production
  • Validate with known binders first

Ligand Preparation

  • Generate 3D structures if starting from SMILES
  • Optimize geometry before docking
  • Calculate Gasteiger charges
  • Set rotatable bonds correctly

Result Analysis

  • Consider both score and RMSD clustering
  • Analyze protein-ligand interactions
  • Visual inspection of binding poses
  • Compare with known binding modes