Skip to main content
. 2020 Jan 20;10:727. doi: 10.1038/s41598-019-56878-x

Figure 2.

Figure 2

Proteome shifts and sensitivities to hypoxia dependent on the metastatic phenotype of the OS cells. (A) Score plot from the principal component analysis (PCA) of the hypoxia responsive proteins (p-value < 0.05 for Treatment term in the LR model) plots; (B) Heat map shows hypoxia-responsive proteins with statistically significant changes in expression levels after LR test. Columns indicate the biological samples, and rows indicate proteins. Protein intensities were normalized by z-scores, which were color-coded as indicated. Hierarchical clustering was performed using Euclidean distances between biological samples. Data used for the generation of the heat map and cluster analysis are compiled in the Supplemental Table S5. Abbreviations used: hh, Hypo-HMPOS; hp, Hypo-POS; nh, Norm-HMPOS; np, Norm-POS. (C) Density plot shows the distributions of z-score subtraction (|zHMPOSzPOS|), which indicates the degree of differences in the expression pattern of hypoxia-responsive proteins between POS and HMPOS cells under hypoxia and normoxia; (D) Density plots show the distributions of the sensitivity scores (|zHypoxiazNormoxia|) for proteins of POS and HMPOS cells, suggesting differential sensitivities of POS and HMPOS cells in response to hypoxic tensions; normalized z-scores of hypoxia-responsive protein targets were calculated based on PRM assay data using extracted ion chromatograms for the proteins and considering all four groups.