Fig. 3.

TAPINTO identified a subgroup with high expression of HER2 based on continuous SD values. (A) Original HER2 expression levels (FPKM) of the 1,109 samples ranked from high to low expression (gray bars). Based on these data, the continuous SD values (blue line) were calculated by iteratively including samples from the highest to the lowest expression. The SD increases as high-expression outliers are included and reaches a maximum at the point where additional lower-expression samples begin to reduce the overall dispersion of the cumulative subset. This maximum value (red dashed line) represents the turning point that objectively defines the boundary between the high-expression subgroup and the normal-like tumor subgroup. (B) The HER2 clinical IHC status (black circles) of 731 of the 1,109 breast cancer samples was either positive (P) or negative (N), as shown, and the results are ordered as above. The standard deviation values of 731 breast cancer samples were also calculated and are shown as the blue line, and the maximum value of the standard deviations is indicated by the red dashed line.