ReCellTy: Domain-specific knowledge graph retrieval-augmented LLMs workflow for single-cell annotation

To enable precise and fully automated cell type annotation with large language models (LLMs), we developed a graph structured feature marker database to retrieve entities linked to differential genes for cell reconstruction. We further designed a multi task workflow to optimize the annotation process. Compared to general purpose LLMs, our method improves human evaluation scores by up to 0.21 and semantic similarity by 6.1% across 11 tissue types, while more closely aligning with the cognitive logic of manual annotation.

@article{han2025_2505.00017,
  title={ ReCellTy: Domain-specific knowledge graph retrieval-augmented LLMs workflow for single-cell annotation },
  author={ Dezheng Han and Yibin Jia and Ruxiao Chen and Wenjie Han and Shuaishuai Guo and Jianbo Wang },
  journal={arXiv preprint arXiv:2505.00017},
  year={ 2025 }
}