Fig. 2

Dynamic network biomarker analyses reveal the tipping point of AST and its regulatory signaling. a Schematic diagram of a theoretical model for system state transition during AST. As lung tumors phenotypically switched from adenocarcinoma (ADC) to squamous carcinoma (SCC), three general phases could be observed: ADC phase, critical phase, SCC phase. The tumor state first gradually changes during the ADC phase, then reaches the tipping point (TP) before the drastic changes, and finally reaches a second slow-changing SCC state. b A composite index (CI) for quantifying the tipping point of system state. The peak of CI at 7 W indicates the tipping point. c. Network dynamics along the sampling time points during the AST process. A key module, called the dynamic network biomarker (DNB), is detected as the critical state at the 7 W (tipping point). Edges were colored by normalized absolute Pearson’s correlation coefficient (PCC). Nodes were colored by individual CI scores. d KEGG pathway enrichment for the DNB genes. Top 5 most significant pathways were shown. e TF enrichment in the upstream of genes repressed across the tipping point. Top 20 TFs are ranked with −log10(P value) and the ADC lineage-specific TFs and Wnt signaling related TFs were highlighted. f Representative photos for immunohistochemical staining of 8-oxo-dGuo and β-catenin in KL mouse lung tumors at 7 W post Ad-Cre treatment. Scale bar: 50 µm. g Correlation of immunohistochemical staining of 8-oxo-dGuo and β-catenin in KL mouse lung tumors at 7 W post Ad-Cre administration. Pearson correlation coefficent (PCC) and P value were shown. h The KL lung cancer cells were treated with indicated ROS inducers (PHEN at 500 μM, PL at 12.5 μM, PEITC at 15 μM) for 6 h and the relative expression of the Wnt pathway-related genes were assessed by real-time PCR. *P < 0.05. **P < 0.01. ***P < 0.001, ****P < 0.0001