Fig. 7: stClinic improves the detection of finer structure by integrating spatial multi-omics data from the same and different slices. | Nature Communications

Fig. 7: stClinic improves the detection of finer structure by integrating spatial multi-omics data from the same and different slices.

From: stClinic dissects clinically relevant niches by integrating spatial multi-slice multi-omics data in dynamic graphs

Fig. 7

a stClinic learns joint features by integrating latent features from multi-omics tools like MultiVi alongside spatial location data within dynamic graphs. b Leveraging aligned features from multi-omics tools like Seurat in a multi-slice integrative condition, stClinic employs the same strategy as to learn the final features. c Manual annotation of mouse brain coronal section, and spatial domains identified by single modality (RNA or ATAC), CellCharter, and stClinic. d Spatial ATAC levels of marker genes for cluster 1 (Pde10a), cluster 3 (Cux2), cluster 8 (Mbp), cluster 12 (Dlx1), and cluster 13 (Drd3), for mouse brain coronal section data denoised by stClinic. e Cell-type distribution of brain, connective tissue, head mesenchyme, heart, liver, lung primordium, mesenchyme, spinal cord, and surface ectoderm on mouse embryo E10.5 tissue profiled by Stereo-seq. f Transferred cell-type distribution on mouse embryo E11 tissue profiled by spatial ATAC-seq using MaxFuse, GLUE, SLAT, Seurat, and stClinic, respectively. g Spatial ATAC levels of marker genes for brain (Sox1ot), spinal cord (Cntnap5b), connective tissue (Dnm3os), and heart (Tnn2), for mouse embryo E11 data denoised by stClinic. Source data are provided as a Source Data file.

Back to article page