Fig. 6

Heterogeneity revealed by Kim dataset. a Pseudo-time ordering reconstruction of the different subgroups: single cells from PDX primary tumor (green), patient metastasis (blue) and PDX metastasis (red). The eeSNVs are obtained with a = 0.6. The tree is inferred using the MST algorithm on Pearson’s correlation-based distance matrix. b Heatmap of the cells (row) and eeSNVs (column). c Bipartite graph using KEGG pathways (orange color) and genes enriched with significant eeSNVs (green color) as two set of nodes. The significant eeSNVs are inferred from the metastasized cells, compared to the primary tumor cells. The size of the nodes reflects the gene scores (given by SSrGE) and the pathway scores (sum of the gene scores). Lighter green indicates genes with a lower rank