Fig. 7: Correlation between JSD, plasticity, and number of positive feedback loops.
From: Identifying inhibitors of epithelial–mesenchymal plasticity using a network topology-based approach
a PS2 for every randomized network for OCT4 module plotted as a function of its JSD from the WT network and number of positive feedback loops. Colorbar shows PS2 scores. b Boxplot for PS2 scores vs. range of JSD values for all networks plotted in a. p-value corresponds to one-way ANOVA test. c, d Same as a, b but for NRF2 module. e For each EMP module, all randomized networks are sub-categorized based on number of positive feedback loops. Each cell denotes correlation coefficient between JSD from WT and plasticity score fold change, wherever applicable (i.e. number of samples corresponding to that network and feedback loop count >3. The white boxes (NaN) represent the cases with number of samples <3). Spearman correlation coefficient shown using color bar; significance represented as : *p < 0.05, **p < 0.01, ***p < 0.001. f For each EMP module, all randomized networks are classified based on range of JSD values. Each cell denotes correlation coefficient between plasticity score fold change and number of positive feedback loops. g, h Same as e, f but for perturbed networks of each EMP module.
