Documentation/Nodes/Molecular Docking

Molecular Docking Nodes

Complete molecular docking workflow with AutoDock Vina integration, GPU acceleration, and comprehensive analysis tools for drug discovery.

Node Reference

Detailed documentation for each molecular docking node available in Bioshift.

AutoDock Vina

CPU-based molecular docking using AutoDock Vina with batch processing support

Type: autodock_vinaCategory: Docking Engine

Key Features

•Batch docking support for multiple ligands
•Automatic RDKit to PDB to PDBQT conversion
•Automatic receptor rigidification for Vina compatibility
•Multiple Vina versions with automatic resolution
•Grid parameters can be provided via input or properties
•Aggregated results across all ligands with naming
•Concatenated logs for debugging and analysis
•Hydrogen management and conformer processing

Input Ports

receptor_moleculesmolecules

Receptor protein structure (RDKit mol)

ligand_moleculesmolecules

Ligand molecules for docking (RDKit mol list)

grid_paramsdata

Optional grid parameters dictionary

Output Ports

docking_resultsdata

DataFrame with docking scores, RMSD values, and ligand names

docked_moleculesmolecules

All poses as RDKit molecules

best_posemolecules

Best scoring pose as single molecule

docking_logstring

Raw docking log text for all ligands

Properties

Property	Type	Default	Description
center_x	float	0.0	Grid center X coordinate (Angstroms)
center_y	float	0.0	Grid center Y coordinate (Angstroms)
center_z	float	0.0	Grid center Z coordinate (Angstroms)
size_x	float	20.0	Grid size X dimension (Angstroms)
size_y	float	20.0	Grid size Y dimension (Angstroms)
size_z	float	20.0	Grid size Z dimension (Angstroms)
exhaustiveness	int	8	Search exhaustiveness (higher = more thorough)
num_modes	int	9	Number of binding modes to generate
energy_range	float	3.0	Energy range cutoff for poses (kcal/mol)
vina_version	string	1.2.5	Vina version to use (1.1.2, 1.2.3-1.2.7)

AutoDock Vina GPU

GPU-accelerated molecular docking using AutoDock Vina GPU with CUDA optimization

Type: autodock_vina_gpuCategory: Docking Engine

Key Features

•CUDA acceleration for high-throughput virtual screening
•Automatic GPU executable resolution
•Batch processing optimization for GPU
•Thread configuration for performance tuning
•Same interface as CPU version for easy switching
•Optimized for NVIDIA GPUs with CUDA support
•Maintained compatibility with standard Vina parameters

Input Ports

receptor_moleculesmolecules

Receptor protein structure (RDKit mol)

ligand_moleculesmolecules

Ligand molecules for docking (RDKit mol list)

grid_paramsdata

Optional grid parameters dictionary

Output Ports

docking_resultsdata

DataFrame with docking scores, RMSD values, and ligand names

docked_moleculesmolecules

All poses as RDKit molecules

best_posemolecules

Best scoring pose as single molecule

docking_logstring

Raw docking log text for all ligands

Properties

Property	Type	Default	Description
center_x	float	0.0	Grid center X coordinate (Angstroms)
center_y	float	0.0	Grid center Y coordinate (Angstroms)
center_z	float	0.0	Grid center Z coordinate (Angstroms)
size_x	float	20.0	Grid size X dimension (Angstroms)
size_y	float	20.0	Grid size Y dimension (Angstroms)
size_z	float	20.0	Grid size Z dimension (Angstroms)
exhaustiveness	int	8	Search exhaustiveness (inherited from CPU version)
num_modes	int	9	Number of binding modes to generate
energy_range	float	3.0	Energy range cutoff for poses (kcal/mol)
gpu_batch	int	50	GPU batch size for processing
thread	int	1000	Thread count per GPU block

Receptor Preparation

Prepare protein receptor for docking (PDB/PDBQT conversion)

Type: receptor_preparationCategory: Preparation

Key Features

•Automatic hydrogen addition
•Water molecule removal
•Missing atom repair
•Format conversion handling

Input Ports

receptor_filefile

Receptor file (PDB/PDBQT)

receptor_moleculesmolecules

Or RDKit molecule

Output Ports

prepared_receptormolecules

Prepared receptor molecule

receptor_pdbqtfile

Receptor PDBQT file path

Properties

Property	Type	Default	Description
add_hydrogens	bool	true	Add missing hydrogens
remove_water	bool	true	Remove water molecules
fix_residues	bool	true	Fix missing atoms in residues

Ligand Preparation

Prepare ligands for docking with charge and torsion assignment

Type: ligand_preparationCategory: Preparation

Key Features

•Gasteiger charge calculation
•Rotatable bond detection
•3D structure generation if needed
•Batch processing support

Input Ports

ligand_filefile

Ligand file (SDF/MOL/MOL2)

ligand_moleculesmolecules

Or RDKit molecules

Output Ports

prepared_ligandsmolecules

Prepared ligand molecules

ligand_pdbqtfile

Ligand PDBQT file paths

Properties

Property	Type	Default	Description
add_hydrogens	bool	true	Add hydrogens
compute_charges	bool	true	Compute Gasteiger charges
detect_torsions	bool	true	Detect rotatable bonds

Grid Search

Define and optimize docking grid box parameters

Type: grid_searchCategory: Grid Setup

Key Features

•Automatic grid centering
•Reference ligand-based positioning
•Active site detection
•Visual grid preview

Input Ports

receptor_moleculesmolecules

Receptor structure

reference_ligandmolecules

Optional reference ligand for centering

Output Ports

grid_paramsdata

Grid parameters dictionary

grid_visualizationimage

Grid box visualization

Properties

Property	Type	Default	Description
method	string	center_of_mass	Grid centering method
padding	float	10.0	Padding around reference
spacing	float	0.375	Grid spacing (Angstroms)

Docking Analysis

Analyze and visualize docking results

Type: docking_analysisCategory: Analysis

Key Features

•Binding affinity analysis
•RMSD-based pose clustering
•Interaction fingerprinting
•Statistical summaries
•Publication-ready plots

Input Ports

docking_resultsdata

Docking results DataFrame

docked_moleculesmolecules

Docked poses

receptor_moleculesmolecules

Receptor structure

Output Ports

analysis_reportdata

Detailed analysis DataFrame

interaction_datadata

Protein-ligand interactions

clustering_resultsdata

Pose clustering analysis

visualizationsimage

Analysis plots

Properties

Property	Type	Default	Description
rmsd_threshold	float	2.0	RMSD clustering threshold
energy_cutoff	float	-6.0	Binding energy cutoff
analyze_interactions	bool	true	Analyze molecular interactions

Vina Split

Split multi-model PDBQT files from docking results

Type: vina_splitCategory: Utilities

Key Features

•Multi-model PDBQT parsing
•Individual pose extraction
•Energy score extraction
•Batch file splitting

Input Ports

pdbqt_filefile

Multi-model PDBQT file

docked_moleculesmolecules

Or docked molecules

Output Ports

split_moleculesmolecules

Individual pose molecules

split_filesfile

Individual PDBQT files

pose_infodata

Pose information DataFrame

Properties

Property	Type	Default	Description
max_poses	int	9	Maximum poses to extract
vina_version	string	auto	Vina version for compatibility

Merge Molecule

Combine receptor and ligand into single complex structure

Type: merge_moleculeCategory: Utilities

Key Features

•Protein-ligand complex creation
•Multi-ligand support
•Format preservation
•Chain ID management

Input Ports

receptor_moleculesmolecules

Receptor structure

ligand_moleculesmolecules

Ligand structure(s)

Output Ports

complex_moleculesmolecules

Merged complex structure

complex_filefile

Complex PDB file

Properties

Property	Type	Default	Description
output_format	string	pdb	Output file format
combine_method	string	append	Molecule combination method

Workflow Examples

Common molecular docking workflows you can build with these nodes.

Basic Molecular Docking

Simple receptor-ligand docking workflow

1Load receptor PDB file using File Input node
2Prepare receptor with Receptor Preparation node
3Load ligand SDF file using SDF Reader node
4Prepare ligands with Ligand Preparation node
5Connect to AutoDock Vina node
6Analyze results with Docking Analysis node
7Visualize with Web 3D Viewer node

High-Throughput Virtual Screening

Screen large compound library with GPU acceleration

1Load receptor structure
2Define grid with Grid Search node
3Load compound library (SDF/CSV)
4Prepare ligands in batch
5Use AutoDock Vina GPU for screening
6Filter by binding energy threshold
7Cluster similar poses
8Export top hits to CSV

Structure-Based Drug Design

Iterative docking with interaction analysis

1Load crystal structure complex
2Extract reference ligand
3Generate grid from reference
4Dock analogs with Vina
5Analyze interactions with ProLIF
6Compare interaction fingerprints
7Visualize binding modes
8Select leads for optimization