Fig. 4: Transcriptional regulation of TNF-α pathway activity along the EMT trajectory.
From: Single-cell multiomic modelling of early metastatic events promoted by the extracellular matrix
a Expression of gradually decreased genes along the EMT trajectory. Rows and columns represent genes and pseudo-ordered cells, respectively. Denoted genes represent TNF-α signaling pathway genes. For visualization, log-normalized expression level of the cells was spline-smoothened and zero-centered with unit variance. b Expression of gradually increased genes along the EMT trajectory. Rows and columns represent genes and pseudo-ordered cells, respectively. Denoted ones represent genes in diverse family such as annexin, S100A, keratin, kallikrein and metallopeptidase. For visualization, log-normalized expression level of the cells was spline-smoothened and zero-centered with unit variance. c (left) Dynamics of the REs of TNF-α signaling pathway genes. Rows and columns represent the REs and pseudo-ordered cells, respectively. For visualization, RE accessibility was zero-centered with unit variance. The REs were grouped into four clusters with hierarchical clustering using Ward.D2 distance. (right) Bar plot represents TF motif enrichment of RE clusters. The pvalue was obtained from HOMER and adjusted with the Benjamini-Hochberg method. d TF motif enrichment along the EMT trajectory. LOESS was used to regress the line, and the Spearman correlation was used to calculate TF motif enrichment according to the pseudo-time trajectory. e Dysregulation of TNF-α signaling genes at the transcriptional and regulatory level. Multimodal profiles (i.e., gene expression and chromatin accessibility) of the cells were zero centered with unit variance. The lines were smoothened with generalized additive models.
