stHGC: a self-supervised graph representation learning for spatial domain recognition with hybrid graph and spatial regularization
Briefings in Bioinformatics · 22.11.2024 · 26 (1)
DOI: 10.1093/bib/bbae666 ↗Kurz gefasst
Fachartikel in der Zeitschrift Briefings in Bioinformatics, online erschienen am 22.12.2024. Verfasst von Runqing Wang und 3 weiteren Autor:innen. Beteiligte Einrichtung: College of Computer Science and Engineering, Dalian Minzu University, 116600 Dalian, China. Hauptthemen laut MEDLINE: Algorithms. Die Arbeit wurde bisher 17-mal zitiert (Crossref) und nennt 45 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: National Natural Science Foundation of China, Natural Science Foundation Project of Liaoning Province, Basic Scientific Research Project of Liaoning Provincial Department of Education.
Abstract (Originalsprache)
Advancements in spatial transcriptomics (ST) technology have enabled the analysis of gene expression while preserving cellular spatial information, greatly enhancing our understanding of cellular interactions within tissues. Accurate identification of spatial domains is crucial for comprehending tissue organization. However, the effective integration of spatial location and gene expression still faces significant challenges. To address this challenge, we propose a novel self-supervised graph representation learning framework named stHGC for identifying spatial domains. Firstly, a hybrid neighbor graph is constructed by integrating different similarity metrics to represent spatial proximity and high-dimensional gene expression features. Secondly, a self-supervised graph representation learning framework is introduced to learn the representation of spots in ST data. Within this framework, the graph attention mechanism is utilized to characterize relationships between adjacent spots, and the self-supervised method ensures distinct representations for non-neighboring spots. Lastly, a spatial regularization constraint is employed to enable the model to retain the structural information of spatial neighbors. Experimental results demonstrate that stHGC outperforms state-of-the-art methods in identifying spatial domains across ST datasets with different resolutions. Furthermore, stHGC has been proven to be beneficial for downstream tasks such as denoising and trajectory inference, showcasing its scalability in handling ST data.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- College of Computer Science and Engineering, Dalian Minzu University, 116600 Dalian, China.
- SEAC Key Laboratory of Big Data Applied Technology, Dalian Minzu University, 116600 Dalian, China.
- Institute of Fundamental and Frontier Sciences, University of Electronic Science and Technology of China, 611730 Chengdu, China.
ORCID-Profile: Runqing Wang, Qiguo Dai, Quan Zou
Förderung
- National Natural Science Foundation of China
- Natural Science Foundation Project of Liaoning Province
- Basic Scientific Research Project of Liaoning Provincial Department of Education
Zitieren
APA
Wang, R., Dai, Q., Duan, X., & Zou, Q. (2024). stHGC: a self-supervised graph representation learning for spatial domain recognition with hybrid graph and spatial regularization. Briefings in Bioinformatics, 26(1). https://doi.org/10.1093/bib/bbae666
Vancouver
Wang R, Dai Q, Duan X, Zou Q. stHGC: a self-supervised graph representation learning for spatial domain recognition with hybrid graph and spatial regularization. Briefings in Bioinformatics. 2024;26(1). doi:10.1093/bib/bbae666
BibTeX
@article{Wang2024,
author = {Wang, Runqing and Dai, Qiguo and Duan, Xiaodong and Zou, Quan},
title = {stHGC: a self-supervised graph representation learning for spatial domain recognition with hybrid graph and spatial regularization},
journal = {Briefings in Bioinformatics},
year = {2024},
volume = {26},
number = {1},
doi = {10.1093/bib/bbae666}
}
EUVIMED