PMC

The percentage of positive genes in 25 top clusters identified by network as described early when eight gene association methods were applied to Arabidopsis microarray data sets of 108 samples.

The percentage of positive genes in the 25 top clusters derived from human data from stem cells underwent differentiation by coexpression network construction and decomposition as described when eight gene association methods were used.

The number of positive regulators in the top 25 TF clusters identified by network construction and decomposition using TF-Cluster when eight gene association methods were applied to the human microarray compendium data set containing 189 chips.

7A, Clusters verse biological events in Arabidopsis: RHG: root hair growth, RCD: root cap growth, RVC: root vascular development, RCC: root cell cycle, DSR: drought response to ABA. 7B. Clusters verse biological events in human: PLM: pluripotency maintenance, ND: neural development; MDD: multi-direction differentiation.

Relative proportions of coexpressed gene pairs are within the same pathways (S); different pathways (D); and none of existing known pathways (N) in the top 100 (left panel) and 500 (right panel) pairs resulting from the correlation analysis. 576 genes in 30 pathways were analyzed against all genes in the genome.

The percentage of genes with approximately a normal distribution of the top 500 genes most tightly coexpressed to each of 9 selected genes as examined with Shapiro-Wilk test (Significance level p<0.01). Pairwise analysis was performed between each of 9 genes and all genes in the genome, and the results were sorted by p values before top 500 genes were cut-off.

Distribution of top 100 genes most closely associated with NANOG when eight gene association methods were applied for pairwise analysis. This analysis was done with 189 human microarray data sets as inputs. The approximate normality of NANOG is shown at the right bottom corner by Q-Q plot, in which the points fall on the reference line (solid line at 45°). Additionally, Shapiro-Wilk test shows NANOG has a W statistic of 0.98.

The numbers of positive TFs in top 25 clusters identified by TF-Cluster when eight gene association methods were used to construct the coexpression network of all TFs (1640 in Arabidopsis ATH1 platform) for network decomposition to recognize the positive TF clusters regulating different biological processes. These TFs were from Nie at al in where literature evidence that support them to be positive genes are shown.

Within and across-pathway gene coexpression connectivity patterns when eight gene associated methods were used. Coexpression relationships are viewed as a heatmap between any two of 9 selected pathways that represent stress response and primary metabolism, and wood formation, which are labeled from 1 to 9. 1) abscisic acid biosynthesis (4 genes), 2) aerobic respiration (14 genes), 3) gluconeogenesis (16 gene), 4) glycolysis IV (plant cytosol) (4 genes), 5) glyoxylate cycle (10 genes), 6) IAA biosynthesis I (10 genes), 7) Phenylpropanoid biosynthesis (10 gene), 8) UDP-galactose biosynthesis (5 genes), 9) UDP-D-xylose biosynthesis (7 genes). For each of 81 genes in these 9 pathways, we obtained the top 100 most coexpressed genes to it by performing a genome-wide coexpression analysis in which each of these 81 genes was paired with all other genes in the genome. We plotted all those pairs, in which both genes are one of these 81 genes in any of these 9 pathways.