Figure 1

Network inference and MRA flowchart. An overview of the bioinformatics pipeline. Input data are 197 Affymetrix.CEL files (63 IBC samples and 134 nIBC samples). The gene expression datasets were analyzed simultaneously with the ARACNe algorithm to infer a transcriptional regulatory network. Master Regulator Analysis was used to select the TFs showing a significant overlap between the targets in each extracted TF regulon in the network and the IBC gene expression signature. ARACNe comprises two main steps: estimation of mutual information between TFs and potential targets and data processing inequality (DPI) to cut most of the indirect interactions.