Fig. 9.
Correlation network graph, subnetworks and network metrics. A, Single connected correlation network graph based on spearman`s correlation value between levels of transcripts, proteins, primary metabolites, secondary metabolites and lipophilic compounds across leaf and stem wild-type and mutant samples. Vertices correspond to individual transcripts, proteins or compounds and are displayed by filled points, colored according to their vertex class. Vertices are connected through edges colored by the sign of the correlation. Graph layout was generated by a progressive multidimensional-scaling algorithm, using the inverse of the edge-weights. Correlation-value threshold = 0.95, nvertex = 12478; nedge = 326539. B, Betweenness, closeness and degree of SCO2 transcript in relation to the overall network. All boxplots show the interquartile range (IQR) between the first and the third quartile as the box, with the median indicated by a black center line. Whiskers extend from quartiles to most extreme points with a maximum of 1.5 x IQR. Points beyond that range are considered outliers but overlaid by individual data points, shown as quasirandom. The respective SCO2 value is indicated by a red cross-circle. Y-axis is transformed by log(1 +x) to accommodate 0 values as well as large outliers. C. Correlation network graph of direct SCO2 neighbors at correlation value threshold 0.95. D, Correlation network graph of photosystem assembly factor genes and proteins at correlation value threshold of 0.75. In C + D, vertices correspond to individual transcripts or proteins and are displayed by filled points, colored according to their vertex class. Vertices are connected through edges colored by the sign of the correlation. In D the outline of the points (stroke) is colored according to the functional grouping. Graph layout was generated by a circular layout algorithm for C and a progressive multidimensional-scaling algorithm for D, using the inverse of the edge-weights. Statistical significance was estimated by single-wilcox test for B. Significant differences are indicated by asterisks for B (*: p ≤ 0.05; **: p ≤ 0.005; ***: p ≤ 0.0005).