Extended Data Fig. 2: Transcriptional rewiring of HSPCs in the BM during breast cancer.
a, Heatmap showing normalized expression of the top 10 most variable genes for each cluster (see Fig. 1H). b, Violin plots showing expression distributions of Meis1, Myl10, Gata1, Elane, Mpo, Dntt, Mki67, Top2a. c, UMAP showing the expression of HSC, MPP2, MPP3 and MPP4 gene signatures from Wilson et al., 2008 and Pietras et al., 2015. Gradient shows the background corrected mean expression of all genes in the signature, calculated using AddModuleScore function. d, Violin plots showing expression distribution of the PyMT response signature (upregulated DEG between PyMT and WT LSK) across clusters (see Fig. 1h). e, Frequency of WT and PyMT-derived cells in each cluster. f, Gating strategy to define LSK cell subsets by FACS. g, LSK cell subsets frequency in tumor-free (n = 9) or Py230 tumor-bearing mice (n = 9, pool of 2 experiments) and Neu mice (n = 6, one experiment). h, Frequencies and cell number of quiescent (TMRELow) and active (TMREHigh) HSCsLT in tumor-free (n = 3) or Py230 tumor-bearing mice (n = 3, one experiment), measured by FACS using TMRE staining. i, UMAP showing isolation of WT and PyMT G2M/S cycling cells from LSK scRNAseq dataset. Cell cycle was assigned based on gene expression profiles using the cyclone method in the scran package. j, UMAP showing the expression of MPP2, MPP3 and MPP4 gene signatures from Pietras et al., 2015 among cycling cells. k, Cell number-normalized density plot showing the cell distribution across the cycling cell dataset. l, Bone marrow cells were enriched for hematopoietic progenitor and stained with EdU to assess cell cycling activity. Frequency of EdU+LSK cell subsets by FACS (n = 6 WT and. 6 PyMT mice, pool of 2 experiments). Data are the mean ± s.d. P value < 0.05 are depicted and were obtained using two-tailed t-test (g, h, and l).
