Figure 4
From: The Notch1 signaling pathway directly modulates the human RANKL-induced osteoclastogenesis
Cell–cell communication network in RANKL-induced osteoclastogenesis. (A) Circos-view of interaction network constructed by NATMI1 algorithm. Edges are colored by ligand expressing cell type, lanes connect target cell type and thickness is relative to the number of ligand–receptor pairs connecting the two clusters. (B) Heatmap of interaction network as reported in A using the NATMI1 method (https://hub.docker.com/r/asrhou/natmi). Cell subsets expressing the receptors are listed in rows. Columns indicate cells expressing the receptors. Color is relative to the count number of ligand–receptor pairs. (C,D) Circos-view and heatmap of interaction network of MC-#A cells using the NATMI1 method (https://hub.docker.com/r/asrhou/natmi). (E,F) mRNA expression level of human genes encoding the cell receptors involved in WNT (E) and JACK-STAT (F) signaling pathways as reported in the KEGG database at the following link https://www.gsea-msigdb.org/. In the plots, read counts, normalized by median ratio of total counts over all genes are reported for the indicated MetaClusters (MC) by SCANPY toolkit (Version: 1.8.1) at the following link https://scanpy.readthedocs.io/en/stable/generated/scanpy.pl.rank_genes_groups_dotplot.html. (G,H) UMAP plots based on the single cell RNA-Sequencing (scRNA-Seq) data showing the expression of RYK receptor and WNT5a ligand (G) as well as the level of IL7R receptor and IL7 ligand (H) in the highlighted MetaClusters (MC) of osteoclast differentiation. (I,J) Visualization of cell types connected based only on WNT5a/RYK (I) or IL7/IL7R (L) matrix (ligand-receptors) by NATMI algorithm. The thickness of lanes is relative to the number of ligand–receptor pairs connecting the two clusters.
