RAG-Enhanced Collaborative LLM Agents for Drug Discovery

2025-02-22

AI Summary

Large language models (LLMs) face significant challenges in drug discovery due to the high heterogeneity and semantic ambiguity of biochemical data, coupled with a scarcity of high-quality annotations; domain-specific fine-tuning is prohibitively expensive, hindering real-time scientific data utilization. To address this, we propose CLADD—a novel multi-LLM collaborative RAG agent architecture specifically designed for biochemical data. CLADD enables molecule-level semantic understanding and reasoning without domain fine-tuning, leveraging dynamic retrieval from biomedical knowledge bases, molecular contextual modeling, and zero-shot cross-task generalization. It effectively resolves key challenges including data heterogeneity, semantic ambiguity, and multi-source evidence integration. Empirical evaluation demonstrates that CLADD significantly outperforms general-purpose LLMs, fine-tuned LLMs, and conventional deep learning models on critical tasks such as target prediction, de novo molecular generation, and ADMET assessment. The system exhibits exceptional flexibility, robustness, and reasoning capability.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Linked Open Data, Knowledge Graphs & KB Completion
• Cognitive Modeling & Cognitive Systems: Computational Creativity

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Recent advances in large language models (LLMs) have shown great potential to accelerate drug discovery. However, the specialized nature of biochemical data often necessitates costly domain-specific fine-tuning, posing critical challenges. First, it hinders the application of more flexible general-purpose LLMs in cutting-edge drug discovery tasks. More importantly, it impedes the rapid integration of the vast amounts of scientific data continuously generated through experiments and research. To investigate these challenges, we propose CLADD, a retrieval-augmented generation (RAG)-empowered agentic system tailored to drug discovery tasks. Through the collaboration of multiple LLM agents, CLADD dynamically retrieves information from biomedical knowledge bases, contextualizes query molecules, and integrates relevant evidence to generate responses -- all without the need for domain-specific fine-tuning. Crucially, we tackle key obstacles in applying RAG workflows to biochemical data, including data heterogeneity, ambiguity, and multi-source integration. We demonstrate the flexibility and effectiveness of this framework across a variety of drug discovery tasks, showing that it outperforms general-purpose and domain-specific LLMs as well as traditional deep learning approaches.

Problem

• Overcoming domain-specific fine-tuning in drug discovery.
• Integrating vast scientific data for drug development.
• Addressing data heterogeneity and ambiguity in biochemical tasks.

Innovation

• RAG-enhanced LLM agents
• Dynamic biomedical data retrieval
• Domain-free drug discovery framework