Figure 3
The multi-NCA biomarker identified when all databases were combined.
Panel A shows the genes and network identified. The connections correspond to data from the PPI database used. The most connected genes were PRKACA, ESR1, LRP1, SMAD2 and JUN. Panel B shows the risk group prediction (splitting the prognostic index by the median) of the multi-NCA biomarker across cancer datasets. Panel C depict the color-coded differential expression of genes between risk groups. Darker red indicates more significant differences. The scales were estimated in -Log10 of the t test p value. Only p values <0.01 are highlighted. Darker purple indicates more significant hazard ratio associations within the Cox model. The scales were expressed in -Log10 of the Z p value. Only p values <0.05 are highlighted. Panel D shows, in the top, the curated biological terms and pathways associated with the genes composing the biomarker. The associations of genes with specific cancers based on the literature are shown at the bottom.
