Fig. 5: Integration of human tonsil CODEX, scRNA-seq and scATAC-seq datasets. | Nature Communications

Fig. 5: Integration of human tonsil CODEX, scRNA-seq and scATAC-seq datasets.

From: scMODAL: a general deep learning framework for comprehensive single-cell multi-omics data alignment with feature links

Fig. 5

a UMAP plots of unintegrated CODEX, scRNA-seq and scATAC-seq datasets, colored by cell types. b UMAP plots of integrated embeddings produced by scMODAL, colored by modalities (left) and cell types (right). c The CODEX section with the original cell-type annotation. d The CODEX section with transferred cell-type annotation. Dashed circles indicate regions showing clear B cell follicle structures. Spatial distributions of B-CD22-CD40 (e) and B-Ki67 cells (f). g Measured CCL4, SLC7A1, DHCR24 and RORA expression levels in the scRNA-seq dataset. Identified spatial distributions of B-CD22-CD40 cells (h) and B-Ki67 cells (i) in the 10x Visium tonsil sample using STitch3D cell-type deconvolution. j Distributions of spatial distances between cells of different cell types and the centers of gravity of germinal centers. k Predicted MKI67 spatial expression pattern in the CODEX dataset. l Measured MKI67 spatial expression pattern in the CODEX dataset. The spatial signaling directions of the CCL4-SLC7A1 and DHCR24-RORA pathways inferred by COMMOT in the CODEX sample (m, n) and the Visium sample (o, p). Dashed circles in (o, p) highlight B cell follicles exhibiting signaling directions consistent with those in the CODEX sample. Source data are provided as a Source Data file.

Back to article page