Fig. 3: ResST identified spatial domains in human breast cancer at a finer level. | Communications Biology

Fig. 3: ResST identified spatial domains in human breast cancer at a finer level.

From: A graph self-supervised residual learning framework for domain identification and data integration of spatial transcriptomics

Fig. 3

a 10× Visium ST data of a breast cancer sample annotated by pathologists containing invasive ductal carcinoma (IDC), ductal/lobular carcinoma (DCIS/LCIS), tumor edge, and healthy regions. b Spatial domains identified by ResST, Seurat, and BayesSpace. c Volcano graph of DEGs between domains 2 and 13 (left); Gene expression with regional annotation and violin plots of PTGES and VTCN1 (right). d Heatmap of Pearson correlation coefficient among domains. e LYZ and SLITRK6 express differentially between domains 1 and 13. f Cell type annotations from Seurat (left); Domains with highest ligand-receptor interaction activity (LR_score < 7000) (middle); Spatial location of domains 5, 7, and 13 (right). g Gene ontology enrichment analysis of the DEGs between domains 1 and 13. h H&E of human breast cancer sample annotated by Agoko’s telepathology platform (left); Spatial domains identified by ResST, SEDR, CCST, and stLearn (right).

Back to article page