Fig. 4

MPE-derived breast tumor in-vitro models reveal heterogeneous drug responses caused by molecular diversity. A See also Table S10 A. A heatmap of MPE-derived breast tumor in-vitro models exhibited heterogeneous distribution of 25 compounds according to their molecular characteristics. The names of compounds are provided on the right. The cell lines and drugs were k-means clustered based on the AUC values across the drug panel. Multiple factors that potentially contribute to the heterogeneous drug responses are indicated above the heatmap. The normality of each drug response was estimated with Shapiro–Wilk test. The p-values < 0.05 was considered as normally distributed data. B See also Table S10B, C. Gene-drug interaction analysis using Wilcoxon ranked sum test. Each dot indicated a pair of gene and drug. The size of circle is proportional to pair count. The absolute log fold change of AUC value > 0.2 and p < 0.05 were considered as significant. C, D. Multiple cell lines reveal heterogeneous drug responses caused by mutational ITH. Shifting mutational landscape during temporal evolution was associated to heterogeneous drug responses (marked with plane figures). The Treeomics statistical settings were as follows (sequencing error rate = 0.005, prior absent probability = 0.5, max absent VAF = 0.05, LOH frequency = 0, false discovery rate = 0.05, false-positive rate = 0.005, and absent classification minimum coverage = 100). E. Longitudinal tracking of mutational fraction revealed drug-associated variant clusters. Each cluster was highlighted with representative colors. Mutational clusters related to varying drug responses were marked with colored dot on the top of the drug heatmap. F. See also Table S11. Correlation plot shows the statistical association of molecular factors contributing to subclonal cell fractions including the VAFs of representative mutations and NES of specific pathways to certain drugs. The Pearson correlation coefficient (R) with p-values between the molecular factors and AUC of six drugs are represented. (Blue: positive correlation; Red: negative correlation). Significance codes: ‘***’ p < 0.001; ‘**’ p < 0.01; ‘*’ p < 0.05