Graph clustering of metabolomic correlated modules in mto1 roots (threshold r ≥ 0.5). Using the DPClus algorithm  we extracted 5 clusters in mto1 roots. The significant metabolic pathways were assigned by KEGG enrichment analysis (see Methods). The central graph consisting of 5 blue clusters and 10 red edges was extracted by DPClus. Each blue cluster contains densely connected metabolites (see Clus1 to 5). Small white nodes in the clusters indicate metabolites. Clus1 (top, left) shows the first cluster of the hierarchical graph. The internal nodes of the clusters are connected by green edges; neighboring clusters are connected by red edges. Both cp and density values, DPClus parameters, were set to 0.5. We used overlapping-mode in DPClus settings. The results obtained in the other 2 genotypes are shown in Additional file 8.
Fukushima et al. BMC Systems Biology 2011 5:1 doi:10.1186/1752-0509-5-1