Fig. 3

Multiple sets of MPE-derived breast tumor cell lines revealed patient-specific transcriptomic alteration. A Principle component analysis (PCA) showed high level of inter-patient heterogeneity. A total of 4 transcriptomic subtypes are identified with PCA. Each subtype is indicated with connected lines and representative colors. B See also Table S5. Multivariate pathway enrichment analysis identifies gene sets that are over-represented across transcriptomic subtypes. Four transcriptomic subtypes identified from PCA are used to estimate pathway aberrations, each indicated by a representative color. Pathways associated with all four subtypes are labeled as “Combined.” The size of each circle corresponds to the pathway term size, representing the number of genes in each pathway. Only pathways with statistically significant enrichment (adjusted p < 0.05) are shown. The filled colors within each circle represent the transcriptomic subtypes contributing to the enrichment of each pathway. C See also Table S6. PCA of Multiple set 3 validates spatiotemporal heterogeneity of MPE-derived breast tumor cell. The principle component 1 (PC1) separates spatially heterogeneous samples with 46.51% of total variation, whereas the principle component 2 (PC2) divides temporally heterogeneous samples with 22.97% of total variation (n = 27,681). Top 0.1% component loadings of PC1 and PC2 were indicated on the top and right side of the PCA plot respectively. D See also Table S7. Pathway enrichment analysis of loading components that account for temporal heterogeneity. The log-transformed p-values are indicated with gradient colors. A value of p <.05 is considered statistically significant. The size of the circle is proportional to the sample size. The fold enrichment is indicated on the x-axis, and the names of enriched pathways are indicated on the y-axis. E Table S8. Normalized enrichment scores (NES) for each Multiple set using single sample gene set enrichment analysis (ssGSEA) display overall temporal transcriptomic changes of MPE samples. Pathways with standard deviation > 0.5 of NES were highlighted with black squares. F EMT and hedgehog signaling pathways are distinctively expressed in the set 2. The statistical settings for GSEA analysis is as follows (Number of permutations = 1000, Permutation type = phenotype, Chip platform = MSigDB.v.7.4.chip, Enrichment statistic = weighted, Max size: exclude larger sets = 500, Min size: exclude smaller sets = 15)