diffdock

Predicts protein-ligand binding poses and confidence scores via molecular docking.

557|98|Updated Nov 7, 2025

Search Tags:#virtual screening #drug discovery #molecular docking #cheminformatics #computational biology #protein-ligand binding

Ready to run

npx skills add https://github.com/jimmc414/Kosmos --skill diffdock-jimmc414

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: diffdock
Source: https://github.com/jimmc414/Kosmos/tree/main/kosmos-claude-scientific-skills/scientific-skills/diffdock
Command: npx skills add https://github.com/jimmc414/Kosmos --skill diffdock-jimmc414

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 (ligands) bind to proteins, a critical step in drug discovery and understanding biological interactions.

Core Features & Use Cases

Molecular Docking: Predicts the 3D binding pose and confidence score of a ligand to a protein target.
Virtual Screening: Enables rapid screening of large compound libraries against a protein target to identify potential drug candidates.
Structure-Based Drug Design: Aids in optimizing lead compounds by predicting their binding modes.
Use Case: A medicinal chemist needs to predict how a new drug candidate might bind to a specific disease-related protein. This Skill can generate the most likely binding poses, guiding further experimental design.

Quick Start

Use the diffdock skill to predict the binding pose of the ligand SMILES '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 drug discovery?▼

Molecular docking predicts 3D binding poses and confidence scores for a ligand and protein target using diffusion-based processing. It accepts standard PDB files or protein sequences to automate binding pose prediction for drug discovery.

Can I run virtual screening of compound libraries against a protein target?▼

Yes, virtual screening supports batch processing of large compound libraries against a specific protein target. It evaluates potential drug candidates by generating binding poses and confidence scores for each molecule.

What input formats do I need for diffusion-based molecular docking?▼

Diffusion-based molecular docking requires ligand SMILES strings and protein structures in PDB format. It also accepts raw protein sequences as input to generate structural complexes for binding pose prediction.

How does structure-based drug design optimize lead compounds?▼

Structure-based drug design optimizes lead compounds by predicting their 3D binding modes to a disease-related protein target. This process generates likely binding poses that guide medicinal chemists in refining experimental designs.

Are there limitations when predicting binding poses for large compound libraries?▼

Predicting binding poses for large compound libraries requires batch processing and careful parameter customization. Users must interpret resulting confidence scores accurately to validate potential drug candidates during virtual screening.