Skip to main content
. 2014 Jun 28;3(5):1112–1125. doi: 10.1002/cam4.277

Figure 1.

Figure 1

OPG DNA copy number and mRNA expression variations in human breast cancer. (A) OPG gene copy numbers in the TCGA-2013 breast cancer set (934 samples) were analyzed at the Oncomine website. The major invasive tumor subtypes display significantly higher OPG DNA copy numbers than normal breast tissue. N, normal breast tissue, ID, invasive ductal carcinoma, IL, invasive lobular carcinoma, IM, invasive mixed carcinoma (ductal and lobular). The P values are calculated with a Student's t-test. (B) The TCGA-2013 breast cancer set was analyzed at the cBioPortal website for the correlation of tumor OPG DNA copy number with patient outcome. A Kaplan–Meier analysis shows a significantly worse prognosis for patients with tumor OPG gene copy gain (red, 182 samples), compared to patients with tumor normal OPG gene copy numbers (blue, 752 samples; P = 4.8 × 10−3). The P value was calculated with a log-rank test. (C) OPG mRNA expression was compared between basal and luminal breast cancer subtypes at the R2 website. The Chin-124 breast cancer set, the largest data set annotated for these subtypes, shows significantly higher OPG levels in basal than in luminal samples (P = 3.4 × 10−4). (D) OPG mRNA expression was compared between basal and luminal breast cancer cell lines at the R2 website. Hoeflich-51, the largest breast cancer cell line data set (51 samples) with this annotation, shows significantly higher OPG levels in basal than in luminal cell lines (P = 1.8 × 10−4). The Barretina-917 cancer cell line set (with 49 breast cancer samples) showed a similar significant difference. In addition, the Garnett-727, Huang-46, and Bild-19 cancer cell lines sets showed the same trend, but without significant P values, probably because of the smaller amount of breast cancer samples in these data sets (38, 21, and 19, respectively). Higher OPG mRNA expression in luminal samples compared to basal samples was never found in any data set (results not shown). The P value in (C and D) was calculated with a Kruskal–Wallis test. In (A–D), the number of samples is in parentheses.