spaMGCN: a graph convolutional network with autoencoder for spatial domain identification using multi-scale adaptation

Spatial domain identification is crucial in spatial transcriptomics analysis. Existing methods excel with continuous and clustered distributions but struggle with discrete ones. We present spaMGCN, an innovative approach specifically designed for identifying spatial domains, especially in discrete t...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Genome Biology Jg. 26; H. 1; S. 159
Hauptverfasser: Zhang, Tianjiao, Zhang, Hongfei, Zhao, Zhongqian, Shao, Saihong, Jiang, Yucai, Zhang, Xiang, Wang, Guohua
Format: Journal Article
Sprache:Englisch
Veröffentlicht: London BioMed Central 10.06.2025
Springer Nature B.V
BMC
Schlagworte:
Animal Genetics and Genomics
Animals
Autoencoder
Bioinformatics
Biomedical and Life Sciences
Cluster analysis
Datasets
Deep learning
Discrete distribution spatial domain
domain
epigenome
Epigenomics - methods
Evolutionary Biology
Gene expression
Gene Expression Profiling - methods
Graphs
Human Genetics
Humans
Identification
Life Sciences
lymph
Lymph nodes
Lymph Nodes - metabolism
Lymphocytes T
Methodology
Mice
Microbial Genetics and Genomics
Multi-source feature fusion
Neighborhoods
Neural networks
Plant Genetics and Genomics
Software
Spatial domain identification
Spatial multi-omics data
Spleen
Spleen - metabolism
Statistical analysis
T-lymphocytes
Transcriptome
Transcriptomics
Discrete distribution spatial domain
Multi-source feature fusion
Spatial domain identification
Spatial multi-omics data
ISSN:1474-760X, 1474-7596, 1474-760X
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Spatial domain identification is crucial in spatial transcriptomics analysis. Existing methods excel with continuous and clustered distributions but struggle with discrete ones. We present spaMGCN, an innovative approach specifically designed for identifying spatial domains, especially in discrete tissue distributions. By integrating spatial transcriptomics and spatial epigenomic data through an autoencoder and a multi-scale adaptive graph convolutional network, spaMGCN outperforms baseline methods. Our evaluations demonstrate its effectiveness in recognizing discrete T cell zones in mouse spleens and follicular cells in human lymph nodes, as well as effectively distinguishing capsule structures from surrounding tissues.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
content type line 23
ISSN:1474-760X
1474-7596
1474-760X
DOI:10.1186/s13059-025-03637-z