Figure 5. Copy number heterogeneity is associated with biological and clinical parameters and tumors can be classified according to clonal composition.

(A) Bar plot with rank order illustrating CNA heterogeneity as measured by fraction (%) of the genome found to be sub-clonal. Each tumor is denoted at the top of each bar with values of fraction (%) genome sub-clonal also annotated. Categories for association analyses are depicted in panel below. Asterisks denote categories with statistically significant associations based on Wilcoxon rank order tests (p-value=0.0229 and 0.0164 for ER- and polyploidy categories respectively). p-values were rounded and annotated in the figure panel. (B) A representative illustration of a Sanger sequencing trace of DNA from a single normal and single cancer cell from tumor P5, an ER+ tumor. Traces illustrate a G to A transition at position 7578500 of chromosome 17 resulting in a non-sense mutation in the TP53 gene. (C) Correlation analysis between fraction (%) of genome sub-clonal and the log of the tumor size (cm2) as measured by MRI. Pearson’s rank correlation p-value=0.01912. Log of the size was used given the assumption of exponential tumor growth. P5, an ER+, TP53 mutant outlier is excluded from the analysis. (D) Representative MRI images of two patient tumors analyzed in this study illustrating the inverse correlation of tumor size and fraction of genome sub-clonal in ER+ tumors. Red colored ellipses denote tumor positions. (E) Heatmap illustration of three representative tumors displaying different clonal composition patterns. Patterns are: homogenous (P16), heterogeneous with a dominant clone (P14), and heterogeneous without a dominant clone (P5). Top bar denotes tumor identity with bottom bar depicting sub-clones. Heatmap color scheme is provided to the right of the heatmap. (F) Quantification and bar plot illustration of clonal composition as explained in E.

Figure 5—figure supplement 1. Copy number heterogeneity is associated with polyploidy and estrogen receptor negative status and these associations are not directly related to the amount of rearrangements found in the cancer genomes of analyzed samples.

(A) Bar plot representation (with rank order) of the number of sub-clonal breakpoints and their association with different biological and clinical parameters (bottom heatmap panel). p-values for associations are rounded and displayed on the figure and are based on Wilcoxon rank order testing. None rounded values are 0.0549 and 0.0241 for ER- and polyploidy categories respectively. (B) Dot plot analysis of the relationship between CNA heterogeneity (quantified using two metrics: fraction (%) genome sub-clonal and the number of sub-clonal breakpoints) and genome rearrangement (number of clonal breakpoints). Not significant relationship is found. (C) Representative illustration of the lack of relationship between sub-clonality (% genome sub-clonal – top bar) and clonal genetic rearrangements (genome-wide copy number profiles). Both tumors are highly re-arranged with one displaying high heterogeneity (P22–53% of genome sub-clonal) while the other displays low heterogeneity (P37–9% subclonal). (D) Bar plot quantification of CNA heterogeneity and patient age of the samples analyzed in this study. No significant relationship is found. (E) Bar plot illustration of the relationship between tumor size and % genome sub-clonal. No significant association is observed when analyzing all tumor biopsies analyzed in the study. A significant-inverse correlation is found when restricting the analysis to ER+ tumors and removing the ER+/TP53 mutant tumor noted in the text. The ER+/TP53 mutant tumor is also denoted in the figure.

Figure 5—figure supplement 2. Clonal composition is variable in breast cancer biopsies and can be used in the classification of tumors.

Heatmap illustration (left panel) and bar plot quantification (right panel) of clonal composition in representative breast tumors illustrating: (A) homogenous tumors with one dominant clone, (B) heterogeneous tumors with multiple sub-clones but with one clone dominant over others (dominant = representative of over 50% of cancer cells), and (C) heterogeneous tumors where many sub-clones exit with none being dominant in composition. Top bar denotes tumor. Bottom bar denotes clone.