Skip to main content
. 2017 Jan 21;8(8):14003–14016. doi: 10.18632/oncotarget.14776

Figure 2. Flow diagram for identifying and validating a prognostic biomarker panel for breast cancer.

Figure 2

The 795 robustly deregulated probe IDs were identified using 4 breast tumor microarray data sets (blue). To identify individual genes associated with relapse-free survival (RFS), Kaplan Meier survival analysis was run on the overlapping IDs (yellow). A gene expression correlation network approach was used to identify cliques of functionally related genes (green). Cox regression was run on 60 random tumor samples for 381 genes significantly associated with RFS (turquoise) to generate the 12-gene signature. The 12-gene signature was used to generate a prognosis scoring system, which was validated using the TCGA and METABRIC BC data sets.