EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

SyRACT: zero-shot biomedical document-level relation extraction with synergistic RAG and CoT

Xin Dong iD, Di Zhao iD, Jiana Meng, Bocheng Guo, Hongfei Lin

Bioinformatics · 01.07.2025 · 41 (7)

DOI: 10.1093/bioinformatics/btaf356 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Bioinformatics, online erschienen am 19.06.2025. Verfasst von Xin Dong und 4 weiteren Autor:innen. Beteiligte Einrichtung: School of Computer Science and Engineering, Dalian Minzu University, Liaoning 116600, China. Hauptthemen laut MEDLINE: Natural Language Processing, Data Mining, Software. Die Arbeit wurde bisher 9-mal zitiert (Crossref) und nennt 32 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Central Universities, Natural Science Foundation of Liaoning Province, China.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Motivation With the advancement of large language models (LLMs), the field of biomedical document-level relation extraction (BioDocRE) has encountered new opportunities. However, LLMs often face challenges such as hallucinated generation, insufficient reasoning capabilities, and a lack of interpretability when performing relation extraction tasks. Results To address these issues, we propose the SyRACT (Synergistic Retrieval Augmented Generation and Chain of Thought) framework for high precision relation extraction in biomedical documents. This framework is built around three core strategies: (i) reframing the relation extraction task as a question answering problem to better align with the processing logic of LLMs; (ii) leveraging an external database constructed from PubMed to provide LLMs with rich and reliable contextual information, thus mitigating hallucination generation; and (iii) construct a specific Chain of Thought for BioDocRE tasks, thereby enhancing the model’s reasoning ability and the interpretability of its output. We validated this approach on three biomedical relation extraction datasets: CDR, GDA, and ADE. Experimental results show that the SyRACT model improves F1 scores by 11.04%, 9.10%, and 41.00% on three datasets, respectively, compared to the DocRE method, which uses standard prompts for LLMs. Availability and implementation Our source code and data are available at https://github.com/donggggxin/SyRACT.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Natural Language ProcessingData MiningSoftwareAlgorithmsHumansDatabases, Factual

Autor:innen und Institutionen

  • School of Computer Science and Engineering, Dalian Minzu University, Liaoning 116600, China.
  • School of Computer Science and Technology, Dalian University of Technology, Liaoning 116024, China.

ORCID-Profile: Xin Dong, Di Zhao

Förderung

  • Central Universities
  • Natural Science Foundation of Liaoning Province, China

Zitieren

APA

Dong, X., Zhao, D., Meng, J., Guo, B., & Lin, H. (2025). SyRACT: zero-shot biomedical document-level relation extraction with synergistic RAG and CoT. Bioinformatics, 41(7). https://doi.org/10.1093/bioinformatics/btaf356

Vancouver

Dong X, Zhao D, Meng J, Guo B, Lin H. SyRACT: zero-shot biomedical document-level relation extraction with synergistic RAG and CoT. Bioinformatics. 2025;41(7). doi:10.1093/bioinformatics/btaf356

BibTeX
@article{Dong2025,
  author = {Dong, Xin and Zhao, Di and Meng, Jiana and Guo, Bocheng and Lin, Hongfei},
  title = {SyRACT: zero-shot biomedical document-level relation extraction with synergistic RAG and CoT},
  journal = {Bioinformatics},
  year = {2025},
  volume = {41},
  number = {7},
  doi = {10.1093/bioinformatics/btaf356}
}
Kostenlos starten →