Verified:GitHub

diffdock

Predict protein-ligand binding poses and confidence scores from PDB or SMILES inputs.

8|Updated Jan 13, 2026

Search Tags:#virtual screening #drug discovery #molecular docking #chemistry #computational chemistry #protein-ligand interaction

Ready to run

npx skills add https://github.com/hxk622/TokenDance --skill diffdock-hxk622

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: diffdock
Source: https://github.com/hxk622/TokenDance/tree/main/backend/app/skills/builtin/scientific/chemistry/diffdock
Command: npx skills add https://github.com/hxk622/TokenDance --skill diffdock-hxk622

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill automates the prediction of how small molecules bind to proteins, a critical step in drug discovery, by identifying potential binding poses and their confidence.

Core Features & Use Cases

Molecular Docking: Predicts the 3D pose of ligands binding to protein targets.
Virtual Screening: Screens large libraries of compounds against a protein target to identify potential drug candidates.
Use Case: A researcher wants to find new potential inhibitors for a specific protein target. They can use this Skill to dock thousands of compounds and identify the most promising ones based on predicted binding poses and confidence scores.

Quick Start

Use the diffdock skill to predict the binding pose of the ligand 'CC(=O)Oc1ccccc1C(=O)O' to the protein structure 'protein.pdb'.

Frequently Asked Questions about diffdock

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I predict protein-ligand binding poses for structure-based drug design?▼

Virtual screening uses diffusion models to dock large libraries of small molecules against a specific protein target, ranking compounds by predicted binding poses and confidence scores to identify promising drug candidates.

How does molecular docking work with diffusion models for computational chemistry?▼

Molecular docking with diffusion models predicts 3D ligand binding poses by iteratively refining molecular orientations against a protein structure, outputting structural conformations and confidence scores for virtual screening campaigns.

Can I use batch processing to screen large compound libraries against a protein target?▼

Yes, batch processing supports virtual screening campaigns by evaluating thousands of SMILES inputs against a protein target, outputting predicted binding poses and confidence scores to identify potential drug candidates.

Do I need PyTorch and RDKit to run protein-ligand docking predictions?▼

Yes, predicting protein-ligand binding poses requires a Python environment with PyTorch, RDKit, and ESM dependencies installed to process PDB structures and SMILES inputs for diffusion model inference.

What is the best way to find new inhibitors for a specific protein target?▼

The best way to find new inhibitors is to run a virtual screening campaign by docking compound libraries against your protein PDB structure, then filtering candidates by their predicted binding poses and confidence scores.

What input formats are supported for predicting protein-ligand interactions?▼

Protein-ligand interaction predictions support PDB files for protein structures and SMILES strings for ligand inputs, enabling both single complex docking and batch virtual screening workflows.