skills.rest
learningmatter-mit
learningmatter-mitOFFICIAL
Explore all skills by this author →
WebsiteGitHub

drug-pocket-detection

Detect and rank ligandable protein pockets using fpocket or P2Rank.

144|21|Updated Jan 8, 2026
Search Tags:#drug discovery#protein#fpocket#p2rank#pocket detection#binding site#docking box
One-click install
npx skills add https://github.com/learningmatter-mit/AtomisticSkills --skill drug-pocket-detection
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: drug-pocket-detection
Source: https://github.com/learningmatter-mit/AtomisticSkills/tree/main/.agents/skills/drug-pocket-detection
Command: npx skills add https://github.com/learningmatter-mit/AtomisticSkills --skill drug-pocket-detection

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill identifies and ranks potential ligandable pockets on a protein structure so you know where drug-like molecules are most likely to bind before running docking.

Core Features & Use Cases

  • Pocket detection on a single conformer: Generates ranked candidate pockets from either geometry-based fpocket (default) or ML-based P2Rank.
  • Per-pocket provenance-rich outputs: Returns geometric center, lining residues, pocket volume (fpocket), and druggability scores (fpocket logistic-regression or P2Rank ligandability probability).
  • Docking-free workflow integration: Explicitly does not dock, but prepares the handoff by converting the chosen pocket into a docking-box input via pocket-to-box scripts.
  • Useful scenarios: When the user provides a protein but has no binding-site information, explores cryptic/allosteric/orphan pockets, or needs to choose where to dock.

Quick Start

Use the drug-pocket-detection skill to run pocket detection on your prepared receptor PDB and return a ranked pocket list with lining residues and druggability scores.

Frequently Asked Questions about drug-pocket-detection

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

FAQPage Schema
How do I detect druggable protein pockets for docking when the binding site is unknown?▼

fpocket detects pockets using a purely geometry-based approach, returning pocket volume and a logistic-regression druggability score. P2Rank uses a machine-learning model to predict ligandability probability, offering a complementary ranking method for the same protein structure.

Can I use pocket detection to identify cryptic or allosteric pockets on apo structures?▼

Yes, pocket detection supports apo and cryptic-pocket exploration. By analyzing the protein's geometric features or using ML predictions, the workflow identifies potential allosteric or orphan sites that may not be visible in holo structures.

What's the best way to convert a detected protein pocket into a docking box?▼

The best way to convert a detected protein pocket into a docking box is to use pocket-to-box scripts that transform the pocket's geometric center and bounding box into the required input coordinates for downstream docking software.

Does this pocket detection workflow actually perform molecular docking on the identified sites?▼

No, this pocket detection workflow explicitly does not perform molecular docking. It is a docking-free workflow designed to identify, rank, and prepare potential binding sites, outputting the necessary geometric data to hand off to a separate docking process.

Related Skills

Explore More Education & Research Skills→
InternScienceInternScience
28

molclaw-quickvina-docking

Performs GPU-accelerated molecular docking and generates pose files with affinity values.

Official
Advanced
jimmc414jimmc414
557

diffdock

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

Community
Advanced
MDheweiMDhewei
64

Protein Structure Alignment

Align two protein structures by Cα superimposition to compute global RMSD and per-residue deviations.

Community
Advanced
InternScienceInternScience
28

molclaw-fpocket-toolkit-base

Detect protein-ligand binding pockets from PDB/mmCIF files and return parsed descriptors and run artifacts.

Official
Intermediate