torchdrug

Automate molecular property prediction, protein modeling, and retrosynthesis with PyTorch GNNs.

Updated Dec 17, 2025

Search Tags:#bioinformatics #drug discovery #graph neural networks #molecular modeling #protein analysis

Ready to run

npx skills add https://github.com/robotlearning123/claude-scientific-skills --skill torchdrug-robotlearning123

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: torchdrug
Source: https://github.com/robotlearning123/claude-scientific-skills/tree/main/scientific-skills/torchdrug
Command: npx skills add https://github.com/robotlearning123/claude-scientific-skills --skill torchdrug-robotlearning123

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires torch, torchdrug, rdkit, tensorflow, networkx, scikit-learn, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill streamlines drug discovery and molecular science research by leveraging state-of-the-art graph neural networks (GNNs) for molecule and protein analysis.

Core Features & Use Cases

Molecular Property Prediction: Predict molecular properties such as solubility, toxicity, and binding affinity.
Protein Modeling: Analyze protein sequences, structures, and properties, including function prediction and structure prediction.
Knowledge Graph Reasoning: Predict missing links and relationships in biological knowledge graphs.
Molecular Generation: Generate novel molecular structures with desired properties.
Retrosynthesis: Plan synthetic routes from target molecules to starting materials.
Use Case: A researcher in drug discovery can use this Skill to predict the ADMET properties of a novel compound and generate potential analogs, streamlining the lead optimization process.

Quick Start

Use the torchdrug skill to predict the activity of a compound with SMILES string 'CCO'.

Frequently Asked Questions about torchdrug

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

FAQPage Schema

How do I predict molecular properties like toxicity and solubility using graph neural networks?▼

Molecular property prediction is automated by this Skill using graph neural networks (GNNs) with PyTorch to calculate attributes like toxicity and solubility. It processes molecular structures directly to output the desired property predictions.

Can I generate novel molecular structures with specific desired properties in PyTorch?▼

Molecular generation in PyTorch is supported to create novel molecular structures with targeted properties. The Skill leverages graph neural networks to synthesize new analog compounds, streamlining the lead optimization process in drug discovery.

Do I need RDKit and PyTorch to perform protein modeling and knowledge graph reasoning?▼

PyTorch and RDKit are required dependencies for this Skill to perform protein modeling and knowledge graph reasoning. These frameworks provide the foundational graph neural network operations and molecular data processing needed for the tasks.

What is the best way to plan synthetic routes for target molecules to starting materials?▼

Retrosynthesis is the best way to plan synthetic routes from target molecules to starting materials. This Skill automates the retrosynthesis process using graph neural networks, allowing researchers to map out chemical synthesis pathways.

Does this GNN approach work for predicting missing links in biological knowledge graphs?▼

Knowledge graph reasoning is supported to predict missing links and relationships in biological knowledge graphs. The Skill applies graph neural networks to infer biological connections, aiding in comprehensive molecular science and bioinformatics research.