Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2507.07426 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912512767688704 |
|---|---|
| author | Yang, Zerui Wan, Yuwei Yan, Siyu Matsuda, Yudai Xie, Tong Hoex, Bram Song, Linqi |
| author_facet | Yang, Zerui Wan, Yuwei Yan, Siyu Matsuda, Yudai Xie, Tong Hoex, Bram Song, Linqi |
| contents | Recent advances in large language models have demonstrated considerable potential in scientific domains such as drug repositioning. However, their effectiveness remains constrained when reasoning extends beyond the knowledge acquired during pretraining. Conventional approaches, such as fine-tuning or retrieval-augmented generation, face limitations in either imposing high computational overhead or failing to fully exploit structured scientific data. To overcome these challenges, we propose DrugMCTS, a novel framework that synergistically integrates RAG, multi-agent collaboration, and Monte Carlo Tree Search for drug repositioning. The framework employs five specialized agents tasked with retrieving and analyzing molecular and protein information, thereby enabling structured and iterative reasoning. Extensive experiments on the DrugBank and KIBA datasets demonstrate that DrugMCTS achieves substantially higher recall and robustness compared to both general-purpose LLMs and deep learning baselines. Our results highlight the importance of structured reasoning, agent-based collaboration, and feedback-driven search mechanisms in advancing LLM applications for drug repositioning. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2507_07426 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | DrugMCTS: a drug repurposing framework combining multi-agent, RAG and Monte Carlo Tree Search Yang, Zerui Wan, Yuwei Yan, Siyu Matsuda, Yudai Xie, Tong Hoex, Bram Song, Linqi Artificial Intelligence Computational Engineering, Finance, and Science Recent advances in large language models have demonstrated considerable potential in scientific domains such as drug repositioning. However, their effectiveness remains constrained when reasoning extends beyond the knowledge acquired during pretraining. Conventional approaches, such as fine-tuning or retrieval-augmented generation, face limitations in either imposing high computational overhead or failing to fully exploit structured scientific data. To overcome these challenges, we propose DrugMCTS, a novel framework that synergistically integrates RAG, multi-agent collaboration, and Monte Carlo Tree Search for drug repositioning. The framework employs five specialized agents tasked with retrieving and analyzing molecular and protein information, thereby enabling structured and iterative reasoning. Extensive experiments on the DrugBank and KIBA datasets demonstrate that DrugMCTS achieves substantially higher recall and robustness compared to both general-purpose LLMs and deep learning baselines. Our results highlight the importance of structured reasoning, agent-based collaboration, and feedback-driven search mechanisms in advancing LLM applications for drug repositioning. |
| title | DrugMCTS: a drug repurposing framework combining multi-agent, RAG and Monte Carlo Tree Search |
| topic | Artificial Intelligence Computational Engineering, Finance, and Science |
| url | https://arxiv.org/abs/2507.07426 |