What inputs are required to run a docking task?

You need a protein target (provided as a PDB file or an amino acid sequence) and a ligand description (provided as a SMILES string or a structure file like SDF or MOL2).

Is a GPU required for this skill?

While DiffDock can run on a CPU, a GPU is strongly recommended as it offers a 10-100x speedup, making it much more practical for complex tasks and batch screening.

Does DiffDock predict binding affinity?

No, DiffDock is specifically designed to predict 3D binding poses and confidence scores. For binding affinity (ΔG), you should combine DiffDock's output with scoring functions like GNINA or MM/GBSA.

How should I interpret the confidence scores?

Scores above 0 indicate high confidence, scores between -1.5 and 0 are moderate, and scores below -1.5 suggest low certainty. High confidence refers to the model's certainty of the pose, not the strength of the binding.

Can I use DiffDock for protein-protein docking?

No, DiffDock is optimized for small molecule ligands. For docking two proteins together, specialized tools like AlphaFold-Multimer or DiffDock-PP are more appropriate.

DiffDock Molecular Docking

Name: DiffDock Molecular Docking
Author: BbgnsurfTech

byBbgnsurfTech

•

データサイエンスとML

Predicts high-accuracy 3D protein-ligand binding poses using state-of-the-art diffusion-based deep learning models.

概要

DiffDock is a specialized Claude Code skill tailored for computational chemists and researchers involved in structure-based drug discovery. By utilizing diffusion models, it predicts the precise 3D spatial orientation of small molecule ligands within protein targets using either PDB structures or amino acid sequences. This skill streamlines the entire molecular docking pipeline, offering automated environment validation, batch processing for virtual screening campaigns, and robust analysis tools to interpret prediction confidence. It is an essential capability for identifying potential binding sites and optimizing lead compounds with high structural accuracy.

主な機能

Provides automated confidence scoring to assess prediction reliability
Handles multiple ligand formats including SMILES, SDF, and MOL2
Includes built-in analysis scripts for ranking and filtering results
Supports batch virtual screening for large-scale compound library processing
Predicts 3D binding poses from PDB structures or ESMFold-supported sequences
3 GitHub stars

ユースケース

Structure-based lead optimization and drug design workflows
Virtual screening of chemical libraries to identify potential drug candidates
Predicting the binding orientation and site of a novel ligand within a protein

概要

主な機能

Provides automated confidence scoring to assess prediction reliability
Handles multiple ligand formats including SMILES, SDF, and MOL2
Includes built-in analysis scripts for ranking and filtering results
Supports batch virtual screening for large-scale compound library processing
Predicts 3D binding poses from PDB structures or ESMFold-supported sequences
3 GitHub stars

ユースケース

Structure-based lead optimization and drug design workflows
Virtual screening of chemical libraries to identify potential drug candidates
Predicting the binding orientation and site of a novel ligand within a protein