virtual-screening

Automate protein-ligand docking and binding site detection workflows.

6|2|Updated Jun 11, 2026

Search Tags:#virtual screening #cheminformatics #AutoDock Vina #SMINA #GNINA #DiffDock-L

Ready to run

npx skills add https://github.com/pradyumnasagar/open-research-skills --skill virtual-screening

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: virtual-screening
Source: https://github.com/pradyumnasagar/open-research-skills/tree/main/skills/chemoinformatics/virtual-screening
Command: npx skills add https://github.com/pradyumnasagar/open-research-skills --skill virtual-screening

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires autodock_vina, smina, gnina, rdkit, meeko, p2rank, openbabel, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill automates virtual screening workflows, enabling efficient identification of potential binders for a given protein target without the need for extensive experimental testing.

Core Features & Use Cases

Docking Tools Integration: Utilizes AutoDock Vina, SMINA, GNINA, and DiffDock-L for various docking scenarios.
Flexible Docking: Offers options for rigid vs. flexible docking, cross-docking vs. self-docking, and binding-site detection.
Receptor and Ligand Preparation: Includes steps for receptor preparation with PDB2PQR, PROPKA, and ligand preparation with meeko and OpenBabel.
Ultralarge Library Screening: Supports screening of ultralarge libraries like ZINC22 and Enamine REAL.
Use Case: When a researcher needs to screen a chemical library against a protein target to identify potential binders for drug discovery.

Quick Start

Run the virtual-screening skill with the following command: virtual-screening -r receptor.pdb -l ligand.sdf --cnn_scoring rescore

Frequently Asked Questions about virtual-screening

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

FAQPage Schema

How do I automate virtual screening workflows for protein-ligand docking?▼

You can automate virtual screening by integrating tools like AutoDock Vina, SMINA, and GNINA to handle receptor and ligand preparation, binding site detection, and docking. This workflow identifies potential binders for a given protein target without requiring extensive experimental testing.

Can I screen ultralarge chemical libraries like ZINC22 using AutoDock Vina and GNINA?▼

Yes, virtual screening workflows support screening ultralarge libraries like ZINC22 and Enamine REAL. The integrated pipeline handles ligand preparation using meeko and OpenBabel, allowing large-scale docking against your target protein using GNINA or SMINA.

What is the difference between rigid and flexible docking in cheminformatics workflows?▼

Rigid docking treats the receptor as a static structure, while flexible docking allows specific receptor side chains to move. This virtual screening workflow offers both options, alongside cross-docking and self-docking scenarios, to accurately predict protein-ligand binding affinities.

Do I need RDKit and OpenBabel for ligand preparation before running SMINA or GNINA?▼

Yes, RDKit, meeko, and OpenBabel are required dependencies for ligand preparation. They convert and optimize chemical structures into the proper formats needed for docking tools like SMINA, GNINA, and AutoDock Vina to process accurately.

How do I detect binding sites for protein targets when I only have a receptor PDB file?▼

Binding site detection is handled using P2Rank within the virtual screening workflow. If you provide a receptor PDB file, the pipeline can automatically identify potential binding pockets before running the docking simulations with AutoDock Vina or GNINA.

Does GNINA support CNN scoring for rescoring docking poses from AutoDock Vina?▼

Yes, GNINA supports convolutional neural network (CNN) scoring for rescoring docking poses. You can execute this by passing the CNN scoring option, such as `--cnn_scoring rescore`, during your virtual screening workflow to refine binding affinity predictions.