Association of signature gene expression with prognosis, breast cancer molecular subtypes, and clinical-pathological features. (a) Heat maps and dendrograms for Wang et al. (panel I), van't Veer et al. (panel II), and Sorlie et al. (panel III) datasets and expression of the 14-3-3ζ signature genes. From left to right, I) analysis of Wang et al. dataset comprising 209 estrogen receptor (ER)-positive breast tumors, II) analysis of 49 ER-positive tumors from the van't Veer et al. study and III) analysis of 67 ER-positive tumors from the Sorlie et al. study. (b) Kaplan-Meier survival curves of patient groups from three independent datasets of ER-positive breast tumors [16–18] based on expression patterns of the 14-3-3ζ gene signature. (c) Subtype classification of tumors with high expression of the signature genes, based on the five breast cancer molecular subtypes. Data were derived from three independent datasets of ER-positive breast tumors in panel a [16–18], as described in Materials and Methods. (d) Association of the 14-3-3ζ gene signature (n = 29 genes) with breast cancer clinical-pathological features. Pearson correlation values are shown. Numbers in bold indicate significant correlations, P < 0.05. NA, not available.