Fig. 6: A predicted gene regulatory network shared across 600 million years of streptophyte evolution.

SWING-RF network filtered based on conserved DPGP clusters of most predictive (enlarged circles) HOGs and metabolites conserved in all three species. All analyses via SWING and DPGP built on the log(fold changes) calculated with the package limma, which statistically models differential gene expression changes. All three biological replicates for each sample were used as inputs per condition. Technically, limma calculates the differences between the mean of groups as the log fold change formula. For Random forest calculations with SWING, the tree number was set to 500. a Predicted conserved stress responsive network including gene expression and metabolite data. Hierarchical orthogroups (HOGs) are annotated if they engage in the top 100 most predictive relationships. Note the colored putative regulatory pathways. Top left, a plot shows the parameters authority and in degree of the hub and non-hub nodes in the network; a scatter shows the data distribution. In all boxplots, horizontal lines within the boxes represent the median (Q2); the box depicts the interquartile range (IQR) from the 25th (Q1) to the 75th percentile (Q3) while the whiskers extend to minimum or maximum values in the data determined by Q1–1.5 × IQR to Q3 + 1.5 × IQR (data points outside are defined as outliers). In all violin plots, we used Gaussian kernel estimation. Tails of the violins were trimmed to the range of the data and violins were scaled to the same area before trimming the tails. b Zoom into the central ‘kinome’ (LRR)–Ser/Thr kinases hub. Inset: domains predicted for the protein encoded by the genes in the Ser/Thr kinases hub. c Section of the via SWING-RF predicted gene regulatory network, re-computed with all time points for which transcriptome data were generated (without metabolite data) centered around the (also there recovered) (LRR)–Ser/Thr kinases hub; for the full network, see Supplementary Fig. 9. d Section of the via SWING-RF predicted gene regulatory network, re-computed with all time points for which transcriptome data were generated (without metabolite data) and filtered based on conserved DPGP clusters of most predictive (enlarged circles) HOGs centered around the (also there recovered) (LRR)–Ser/Thr kinases hub; for the full network, see Supplementary Fig. 10.