Figure 2.

Reconstruction of gene regulatory networks to identify a master regulator. (A) Schematic overview of the workflow. Tumor spheroids derived from the 11 NSCLC cell lines were treated with cisplatin 20 μM for 24 h for RNA sequencing (RNA-seq). Using RNA-seq data, ARACNe and MRA were performed to identify the master regulators and their target genes. Then, 10 potential candidate transcription factors were functionally validated. (B) Heatmap showing the hierarchical clustering of the cisplatin-sensitive and -resistant cell lines and the fold-change values of 518 signature genes. The signature genes were filtered by the ratio of average fold-change in sensitive and resistant groups more than 2. (C) Network visualization of 110 master regulators which was predicted by ARACNe-MRA. The number of nodes and edges is 3,549 and 28,633, respectively. The yellow circles contain 110 master regulators and the red circle indicates top 10 master regulators. (D) List of top 10 master regulators ranked by the order of p-value. Plus mode means that the transcription factors are positively correlated with its target genes. R/S: average fold-change ratio of cisplatin to veh between resistant (R) and sensitive (S) groups.