Fig. 6
From: Inference of single-cell network using mutual information for scRNA-seq data analysis
Detection comparison of cell-type markers in human foreskin fibroblast (HFF) using the SINUM DMs, CSN DMs, and GEMs based on Chu-type dataset. (A) t-SNE plot of GEM for Chu-type dataset, colored by cell types (ARI = 0.89 and FMI = 0.92). The cell type of HFF (green) was chosen to detect the potential markers using the SINUM DMs, CSN DMs, and GEMs, such as gene MTHFD1 (methylenetetrahydrofolate dehydrogenase 1) and gene pair MTHFD1-IFI6 (interferon alpha inducible protein 6). Gene MTHFD1 in the t-SNE plots, colored by (B) the gene expression level and the network degree level in the (C) SINUM SCNs and (D) in the SINUM SCNs. The degree d for networks and raw count r for gene expression were transformed by log10(d + 1) and log2(r + 1), respectively. Fisher's exact test was performed to statistically test whether the proportion of gene MTHFD1 as a hub in SCNs is higher among the HFF cells than among the other cells. The nodes with degrees within the top 25% of all nodes were defined as the hubs (i.e., hub genes) of each SCN. The odds ratio (OR) and p value (P) of statistical analysis are shown. Edge MTHFD1-IFI6 in the t-SNE plots, colored by the (E) SINUM edge scores (i.e., z scores) and (F) CSN edge scores (i.e., z scores). Fisher's exact test was performed to statistically test whether the proportion of edges in SCNs is higher among the HFF cells than among the other cells