EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

scHilda: Hierarchical Integration of LLM with KG database for single cell type annotation

Yilang Li, Yidi Sun, Aoyun Geng, Junlin Xu, Yajie Meng, Feifei Cui, Leyi Wei, Quan Zou iD, Xiulai Li, Zilong Zhang iD

PLOS Computational Biology · 11.05.2026 · 22 (5) · e1014291

DOI: 10.1371/journal.pcbi.1014291 ↗

Kurz gefasst

Fachartikel in der Zeitschrift PLOS Computational Biology, online erschienen am 11.05.2026. Verfasst von Yilang Li und 9 weiteren Autor:innen. Beteiligte Einrichtung: School of Cyberspace Security (School of Cryptology), Hainan University, Haikou, China. Hauptthemen laut MEDLINE: Computational Biology, Single-Cell Analysis, Molecular Sequence Annotation. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 49 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Cell type annotation in single-cell RNA sequencing is a critical bottleneck, with existing automated methods facing limitations in accuracy, interpretability, and generalization to novel cell types. Although Large Language Models (LLMs) have recently shown potential in single-cell annotation, they are prone to inherent “hallucinations”. Furthermore, a critical challenge is utilizing imperfect and potentially noisy external knowledge bases in a principled and robust manner to effectively constrain and enhance the LLM’s reasoning capabilities. To address this, we propose scHilda, a novel framework designed to tackle this challenge. It deeply integrates an external Knowledge Graph into the LLM’s reasoning process and employs a hierarchical arbitration annotation strategy. This strategy first identifies major cell lineages with the support of a global knowledge base and then dynamically retrieves focused subgraph domain information related to that lineage to precisely resolve cell subtypes. This dynamic knowledge-enhanced reasoning mechanism effectively constrains the LLM’s decision space, reduces the risk of hallucination, and mitigates potential misguidance from knowledge base deficiencies. Tests on multiple benchmark datasets show that scHilda outperforms existing methods, achieving state-of-the-art (SOTA) performance. Notably, scHilda demonstrates exceptional robustness when handling complex mixed samples and enables lower-cost lightweight LLMs to achieve annotation performance close to that of top-tier models. Furthermore, rigorous statistical evaluations, alongside detailed interpretability case studies and query complexity analyses, validate the framework’s efficiency and transparent decision-making. By deeply integrating the reasoning power of LLMs with structured biological knowledge, scHilda not only improves the accuracy and interpretability of cell annotation but also provides a new paradigm for building the next generation of trustworthy biological AI systems.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Computational BiologySingle-Cell AnalysisMolecular Sequence AnnotationLarge Language ModelsHumansAnimalsKnowledge BasesCell LineageSoftwareAlgorithms

Autor:innen und Institutionen

  • School of Cyberspace Security (School of Cryptology), Hainan University, Haikou, China.
  • School of Computer Science and Technology, Hainan University, Haikou, China.
  • School of Computer Science and Technology, Wuhan University of Science and Technology, Wuhan, Hubei, China.
  • School of Computer Science and Artificial Intelligence, Wuhan Textile University, Wuhan, Hubei, China.
  • Centre for Artificial Intelligence driven Drug Discovery, Faculty of Applied Science, Macao Polytechnic University, Macao Special Administrative Region, China.
  • Institute of Fundamental and Frontier Sciences, University of Electronic Science and Technology of China, Chengdu, China.

ORCID-Profile: Quan Zou, Zilong Zhang

Förderung

  • National Natural Science Foundation of China

Zitieren

APA

Li, Y., Sun, Y., Geng, A., Xu, J., Meng, Y., Cui, F., Wei, L., Zou, Q., Li, X., & Zhang, Z. (2026). scHilda: Hierarchical Integration of LLM with KG database for single cell type annotation. PLOS Computational Biology, 22(5), e1014291. https://doi.org/10.1371/journal.pcbi.1014291

Vancouver

Li Y, Sun Y, Geng A, Xu J, Meng Y, Cui F, et al. scHilda: Hierarchical Integration of LLM with KG database for single cell type annotation. PLOS Computational Biology. 2026;22(5):e1014291. doi:10.1371/journal.pcbi.1014291

BibTeX
@article{Li2026,
  author = {Li, Yilang and Sun, Yidi and Geng, Aoyun and Xu, Junlin and Meng, Yajie and Cui, Feifei and Wei, Leyi and Zou, Quan and Li, Xiulai and Zhang, Zilong},
  title = {scHilda: Hierarchical Integration of LLM with KG database for single cell type annotation},
  journal = {PLOS Computational Biology},
  year = {2026},
  volume = {22},
  number = {5},
  pages = {e1014291},
  doi = {10.1371/journal.pcbi.1014291}
}
Kostenlos starten →