Development of an Independent Prognostic Signature Based on Three Hypoxia-Related Genes for Breast Cancer

Figure 2

The consensus clustering analysis of BC samples based on the mRNA expression values of 26 HRGs. (a) The elbow chart to determine the optimal number of clustering. The horizontal axis: the numbers of clustering (k) and the vertical axis: the sum of squared error. The point where the decline tending to be flat was the optimal number of clustering (). (b) The consistency matrix of samples and different colors represented different clusters. (c) The heat maps of expression values of HRGs in three clusters. The horizontal axis represented genes; the vertical axis represented samples; red indicated high expression; blue indicated low expression. (d) PCA analysis. The dots with different colors and shapes represented different groups of samples and the closer two dots, the more similar the expression of HRGs in the samples. (e) The Kaplan Meier survival curve of three clusters. value was calculated by log-rank test.

(a)

(b)

(c)

(d)

(e)