Fig. 2

MiBiOmics networks exploration. MiBiOmics networks analysis visualizations (network exploration section) from the analysis of the The Cancer Genome Atlas Network [14] breast cancer TCGA datasets. a Correlation heatmap displaying associations of interest between mRNA’s WGCNA modules and contextual parameters. b The upper panel indicates the contribution of each sample in the red mRNA module delineation. Module eigenvalues are calculated for each sample and indicate how much they participate in the inference of each module. The lower panel indicates the corresponding subtype value for each sample. Here, Basal samples positively contribute to module red of mRNA, while Her2, LumA and LumB negatively contribute to the mRNA red module. c Hive plot displaying the protein red module's features according to their VIP scores, correlations to the subtype parameter and their relationships. In this hive plot, each point represents a variable of the protein WGCNA red module ordered, on the x-axis according to its VIP score, and on the y-axis according to its correlation to the subtype parameter. Edges linking proteins represent the actual edges of the WGCNA network. This representation is useful to distinguish central variables in the module (associated to many other variables) and predictive variables (features with a high VIP), and thus to assert whether central variables of the module are more associated to changes in the discriminant trait compared to predictive variables