skills.rest
jaechang-hits
jaechang-hitsCOMMUNITY
Explore all skills by this author →
GitHub

torchdrug

Train GNN models for molecular property prediction using PyTorch and TorchDrug.

298|27|Updated Feb 18, 2026
Search Tags:#machine learning#drug discovery#pytorch#cheminformatics#graph neural networks#molecular property prediction
One-click install
npx skills add https://github.com/jaechang-hits/SciAgent-Skills --skill torchdrug-jaechang-hits
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: torchdrug
Source: https://github.com/jaechang-hits/SciAgent-Skills/tree/main/skills/structural-biology-drug-discovery/torchdrug
Command: npx skills add https://github.com/jaechang-hits/SciAgent-Skills --skill torchdrug-jaechang-hits

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires torchdrug, torch, torch-geometric, rdkit, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill streamlines complex drug discovery tasks by providing a powerful machine learning platform for molecular representation, property prediction, and interaction modeling.

Core Features & Use Cases

  • Molecular Property Prediction: Train models to predict ADMET properties or bioactivity.
  • Drug-Target Interaction (DTI): Build models to predict binding affinity between compounds and proteins.
  • Retrosynthesis: Predict synthetic routes for target molecules.
  • Use Case: Predict the blood-brain barrier penetration for a new drug candidate using a pre-trained GNN model.

Quick Start

Use the torchdrug skill to train a GNN model for predicting molecular properties 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 ADMET using graph neural networks?▼

Molecular property prediction for ADMET is done by training graph neural network architectures on benchmark datasets. You can train models to predict blood-brain barrier penetration or bioactivity directly from molecular graph representations.

Can I use this to model drug-target interactions and predict binding affinity?▼

Yes, you can build drug-target interaction models to predict binding affinity between compounds and proteins. The platform leverages PyTorch to facilitate interaction modeling using graph-based molecular representations.

Do I need RDKit and PyTorch installed to process molecular graphs?▼

Yes, you need RDKit and PyTorch installed for molecular graph processing and model execution. These dependencies are strictly required alongside TorchDrug to handle cheminformatics data and run graph neural network training.

What is the best way to predict synthetic routes for target molecules in drug discovery?▼

Retrosynthesis prediction enables you to predict synthetic routes for target molecules. By leveraging PyTorch-based machine learning, the platform provides architectures to map out viable synthesis pathways for drug candidates.

Does this approach work for predicting blood-brain barrier penetration with pre-trained models?▼

Yes, you can predict blood-brain barrier penetration for new drug candidates using pre-trained GNN models. It utilizes the BBBP benchmark dataset to evaluate and execute molecular property predictions.

Related Skills

Explore More Education & Research Skills→
jackspacejackspace
16

torchdrug

Apply graph neural networks to drug discovery tasks with PyTorch.

Community
Advanced
InternScienceInternScience
28

molclaw-admet

Predicts ADMET properties for batches of molecules from SMILES strings or files.

Official
Intermediate
nahisahonahisaho
3

scientific-graph-neural-networks

Apply graph neural networks to predict molecular properties, protein interactions, and knowledge-graph relations.

Community
Advanced
sanand0sanand0
8

torchdrug

Predict molecular properties with graph neural networks using PyTorch.

Community
Advanced