Spatial Domain Identification of Spatial Transcriptomics Data for Breast Cancer based on Locally Weighted Ensemble
10.11783/j.issn.1002-3674.2025.04.002
- VernacularTitle:基于局部加权集成的乳腺癌空间转录组数据的空间域识别
- Author:
Hongyan CAO
1
;
Gaiqin LIU
;
Yaxin TIAN
Author Information
1. 山西医科大学卫生统计教研室,重大疾病风险评估山西省重点实验室,煤炭环境致病与防治教育部重点实验室 030001
- Publication Type:Journal Article
- Keywords:
Locally weighted ensemble;
Spatial transcriptomics;
Spatial domain identification;
Breast cancer
- From:
Chinese Journal of Health Statistics
2025;42(4):486-490,495
- CountryChina
- Language:Chinese
-
Abstract:
Objective The locally weighted ensemble based spatial domain identification(LWESDI)method is proposed to explore its application in spatial domain identification in breast cancer spatial transcriptomics data.Methods The LWESDI method is applied to integrate the spatial domain identification results from four methods:BayesSpace,BASS,SpaGCN,and STAGATE,which are used for breast cancer.A locally weighted co-association matrix is constructed by combining the weighted similarity between spots.Obtain a consistent spatial domain identification result by iteratively merging the regions with the highest similarity.Subsequently,differential analysis is performed on the selected highly variable genes,followed by GO enrichment analysis of the differential genes.Results The LWESDI method accurately identifies 20 spatial domains in breast cancer tissue,outperforming the four base clustering methods in terms of accuracy and robustness.The top 3000 highly variable genes(HVGs)were selected,and GO enrichment analysis was performed on the 19 most significantly differentially expressed genes in breast cancer,resulting in 33 enriched GO terms.Conclusion The LWESDI method provides a new strategy for spatial domain identification.The selected potential biomarkers for breast cancer will offer potential therapeutic targets for the study of breast cancer heterogeneity and personalized treatment.
