Fig. 3

The first validation study conducted with the GSE52519 dataset a revealed four clusters for the 57 mutual probesets, the rightmost group in the heat map represents the normal samples (N), two BC samples were located in the same cluster (ST1), four BC samples were clustered in the third group (ST2), and the fourth cluster included three BC samples that clustered separately (ST3); this validation included the CKAP2L and AURKB genes [red arrows] that were upregulated in ST2 and ST3 [BC and control samples are represented on the vertical axis, probes (genes) on the horizontal axis; red, black, and green colors represent high, medium, and low gene expression levels, respectively]. b The Heuristic Online Phenotype Prediction (HOPP), a gene expression analysis algorithm that employs PEARSON (r) correlations, grouped both (normal and BC) phenotypes and predicted two BC samples along with the normal sample