torchdrug

Develop custom GNN architectures for molecular and protein data analysis with PyTorch.

Updated Mar 10, 2026
One-click install
npx skills add https://github.com/Yezez9/Research-Agent --skill torchdrug-yezez9
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: torchdrug
Source: https://github.com/Yezez9/Research-Agent/tree/main/scientific-skills/torchdrug
Command: npx skills add https://github.com/Yezez9/Research-Agent --skill torchdrug-yezez9

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill provides a powerful PyTorch-native toolkit for building custom graph neural network (GNN) architectures to analyze molecules and proteins, accelerating research in drug discovery and molecular science.

Core Features & Use Cases

  • Molecular Property Prediction: Predict properties like solubility, toxicity, and activity.
  • Protein Modeling: Analyze protein sequences and structures for function prediction.
  • Knowledge Graph Reasoning: Perform link prediction on biomedical knowledge graphs for drug repurposing.
  • Molecular Generation & Retrosynthesis: Design novel molecules and plan synthetic routes.
  • Use Case: Predict the binding affinity of a new drug candidate to a target protein using GNNs trained on existing protein-ligand data.

Quick Start

Use the torchdrug skill to predict molecular properties using the GIN model on the BBBP dataset.

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 using graph neural networks in PyTorch?

To predict molecular properties using graph neural networks in PyTorch, you can use this Skill to train models like GIN on datasets such as BBBP for predicting solubility, toxicity, and activity. It integrates with RDKit for cheminformatics.

Can I use graph neural networks for protein function analysis and drug repurposing?

Yes, graph neural networks for protein function analysis and drug repurposing are supported. This Skill enables protein sequence and structure analysis for function prediction, alongside link prediction on biomedical knowledge graphs.

How does PyTorch integrate with RDKit for molecular generation and retrosynthesis?

PyTorch integrates with RDKit for molecular generation and retrosynthesis by providing a PyTorch-native toolkit to build custom graph neural network architectures that design novel molecules and plan synthetic routes.

Do I need PyTorch to build custom GNN architectures for molecular and protein data analysis?

Yes, you need PyTorch to build custom GNN architectures for molecular and protein data analysis. This Skill is a PyTorch-native toolkit that requires the PyTorch framework to develop models for molecular property prediction and protein modeling.

What is the best way to predict binding affinity of a drug candidate to a target protein?

The best way to predict binding affinity of a drug candidate to a target protein is using GNNs trained on existing protein-ligand data. This Skill facilitates developing custom architectures for this exact drug discovery task.

Can I perform knowledge graph reasoning for biomedical applications using GNNs?

Yes, you can perform knowledge graph reasoning for biomedical applications using GNNs. This Skill supports link prediction on biomedical knowledge graphs to facilitate tasks like drug repurposing and molecular interaction analysis.