DDI-GPT addresses the challenge of predicting and understanding drug-drug interactions (DDIs) by leveraging Large Language Models (LLMs) enhanced with Knowledge Graphs (KGs). It targets researchers and clinicians needing explainable DDI predictions, offering a web server for interactive exploration of KG-derived insights and interaction networks. The primary benefit is enhanced interpretability and deeper understanding of potential DDIs beyond simple prediction.

How It Works

The system processes drug pairs identified by DrugBank IDs, generating input sentences enriched with biomedical entities from a KG. A prediction pipeline then analyzes this data, with an explanation module highlighting important words that influenced the outcome. The DDI-GPT web server visualizes these explanations and allows users to explore the underlying KGs, drug-protein-drug interaction networks, and drug-side-effects-drug interaction networks. Users can also visualize shortest or two-hop paths between drug combinations, offering a multi-faceted view of potential interactions.

Quick Start & Requirements

• Primary install/run command: Local execution is initiated via streamlit run Introduction.py.
• Prerequisites: Requires DrugBank IDs for drug input. The system relies on associated KGs and interaction networks for its functionality.
• Online Demo: An interactive walkthrough is available at https://pyvisddi-24u28afk4upfhpvclyujvs.streamlit.app/.
• Setup Time/Footprint: Specific setup time and resource requirements are not detailed in the provided description.

Highlighted Details

• LLM-based DDI prediction augmented with KG data.
• Explainable AI (XAI) features for prediction interpretability.
• Interactive visualization of KGs and diverse drug interaction networks.
• Pathfinding capabilities to illustrate relationships between drug combinations.

Maintenance & Community

No information regarding contributors, community channels, or project roadmap was present in the provided text.

Licensing & Compatibility

The README snippet does not specify a license type or compatibility notes for commercial or closed-source use.

Limitations & Caveats

The provided description does not explicitly state limitations, such as unsupported drug classes or specific KG data requirements. The reliance on external KGs and DrugBank IDs suggests potential data dependency.