diffdock - Predict protein–small-molecule binding poses
Predict 3D binding poses and confidence scores for small molecules docked to proteins from structures or sequences.
Tags
Updated: 2026-10-03Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Predict ligand binding poses
- Process protein PDB files
- Fold protein sequences with ESMFold
- Process SMILES, SDF, and MOL2
- Run batch virtual screening
- Analyze confidence scores
- Rank docking predictions
- Export result summaries
Inputs
- Protein PDB files
- Protein amino acid sequences
- Ligand SMILES strings
- Ligand SDF files
- Ligand MOL2 files
- Batch input CSV files
- Inference configuration files
Outputs
- Ranked ligand pose SDF files
- Pose confidence scores
- Docking result directories
- Prediction analysis summaries
- Exported result CSV files
Requirements
- DiffDock repository
- Python 3.9 environment
- RDKit
- PyTorch and PyTorch Geometric
- ESM
- Read, write, edit, Bash, Glob, and Grep access
- Optional CUDA GPU acceleration