Protein and RNA landscape of PV stem and progenitor cells. (A) Heatmap (unsupervised complete hierarchical clustering of Euclidean distances) for the 3769 proteins uniquely identified across HSC/MPPs and CMP/MEPs isolated from 18 patients with PV and 21 controls. z-Score log2(I) refers to the z-score of log2 transformed protein intensities. Clustering was observed mainly according to cell type. Within cell subpopulations, patients with progressed PV (marked in green and orange) clustered away from patients with chronic PV (marked in blue and purple), which in turn clustered away from control subjects (marked in red). For graphs with separate clustering analyses in HSC/MPPs and CMP/MEPs, see supplemental Figure 5A-B. (B) Heatmap of the 2000 most variant mRNAs in HSC/MPPs, CMP/MEPs, CMPs, MEPs, and GMPs isolated from patients with PV and controls, confirming clustering mainly according to cell type. Class (patient with PV or control), tx (treatment with HU or untreated [UT]), and progression status (chronic [chron] or progressed [prog]) is marked in color. (C) Corresponding principal component analysis of protein intensities. (D) Multidimensional scaling plot of RNA-seq data. (E) Correlation between protein and RNA expression. Gray lines mark the minimum, maximum, and the first and third quartile, and the green line represents the median. Most correlations were positive, but negative correlations were also observed. See supplemental Figure 7 for representative plots of specific candidates with positive and negative correlations and supplemental Figure 8 for an analysis of protein intensity vs RNA intensity. norm, normalized.