Extended Data Fig. 3: Regulons identified by GRN inference. | Nature Cell Biology

Extended Data Fig. 3: Regulons identified by GRN inference.

From: Robust gene expression programs underlie recurrent cell states and phenotype switching in melanoma

Extended Data Fig. 3

a, The top 20 regulons for each state were extracted and AUCell values (Z-score) are plotted in box plots. HES6 regulon for the melanocytic state; NFATC2, EGR3, ETV4, ELF1 and SOX6 regulons for the intermediate state; and FOSL2, JUN regulons for the mesenchymal-like state have distinct AUCell values across the different states. Box plots display the median as centre line and the upper and lower quartiles as box limits (n=100 runs; dots represent the filtered regulons). b, Scatter plots of AUCell values for corresponding motif- and track-based regulons for each cell and the resulting Spearman correlation coefficient, for the 21 recurrently-observed transcription factors (n=4,322). c, Scatter plots of AUCell values for corresponding track- and motif-based regulons for each cell and the resulting Spearman correlation coefficient for the remaining transcription factors (n=4,322). d, Violin plots for motif- and track-based MITF regulon activity are shown for 10 melanoma cultures demonstrating a gradual decrease from extreme melanocytic to intermediate cultures. Violin plots display cell densities as shape and cells as dots (n=4,322).

Source data

Back to article page