Figure 2From: Correlation Network Analysis reveals a sequential reorganization of metabolic and transcriptional states during germination and gene-metabolite relationships in developing seedlings of ArabidopsisRelationships between metabolites. (A) Clusters of metabolites with similar profiles generated by 2D-SOM. Hierarchal and K-means clustering were used to estimate the optimal number of bins for 2D-SOM analysis. Metabolites in cluster 1: sucrose, rhamnose, citrate, alanine, trigonelline, lactate, glucose, threonine, unkS7.37, unkM1.85; Cluster 2: arginine, formate; Cluster 3: fumarate, proline, glutamate, unkD8.0, unkM5.18, unkD3.12; Cluster 4: malate; Cluster 5: valine, isoleucine, leucine, choline, unkD5.69; Cluster 6: fructose, glutamine, unkM7.9. (B) Spring embedding plots showing relationships based on correlations. The plot shows metabolites as nodes and Pearson correlation coefficients over days as connections. The color of the connecting line describes the strength of the correlation between the nodes; a dark red color indicates a strong positive correlation and a dark blue line represents a weaker positive correlation according to the scale of correlation coefficients on the right of the graph. Only correlations above a Bonferroni-adjusted P-value < 0.0001 are shown. (C) Enlargement of the lactate cluster. (D) Enlargement of the valine cluster. Since values start from an initial random configuration, the directions separating cluster in each spring embedding plot are arbitrary, but they provide an indication of distance separating nodes and edges.Back to article page