Figure 1.
Motif enrichment and motif position bias analysis.
(A) A representation of the co-expression network. Each node represents a gene, while a connection between two genes indicates similar expression pattern. Genes containing a given motif are shown in red. Genes in group II are enriched for this motif and are more likely to be regulated by that motif. (B) In this sub-network, both gene groups include similar numbers of genes containing the motif, but the motif positions distribute differently along the promoters in two modules. Group III genes display position bias for the motif and are more likely to be regulated by that motif.
Figure 2.
A sub-network for the G-box motif based on motif enrichment analysis.
Genes were identified by motif enrichment analysis with pValue for the G-box motif < = 0.001. Eleven modules were identified and labeled with the name of the representative gene. Red nodes – genes whose promoters contain the G-box motif; white nodes – genes whose promoters lack the G-box motif.
Table 1.
GO enrichment of co-expression modules identified in the G-Box sub-network*.
Figure 3.
A sub-network for the G-box motif based on the motif position bias analysis.
Genes were identified in the motif position analysis with z-score> = 3. Thirteen modules were identified. Among them, 9 modules (circled in blue) were also identified via the motif enrichment analysis (See Figure 2), while 4 modules (circled in green) represent additional modules identified via the motif position analysis. Red nodes – genes whose promoters contain the G-box motif; white nodes – genes whose promoters lack the G-box motif.
Figure 4.
A sub-network for the MYB motifs based on motif enrichment analysis.
Genes were identified in the motif enrichment analysis with a pValue for MYB motifs < = 0.01. Ten modules were identified, labeled with the name of the representative gene. Red nodes – genes whose promoters contain the MYB motif; white nodes – genes whose promoters lack the MYB motif.
Table 2.
GO enrichment of co-expression modules identified in the MYB motif sub-network*.
Figure 5.
A sub-network for the MYB motifs based on motif position bias analysis.
Genes were identified in the motif position analysis with z-score> = 2.2. Sixteen modules were identified. Among them, 7 modules (circled in blue) were also identified via the motif enrichment analysis (See Figure 4), while 8 modules (circled in green) represent additional modules identified via the motif position analysis. Red nodes – genes whose promoters contain the MYB motif; white nodes – genes whose promoters lack the MYB motif.
Figure 6.
A sub-network for the W-box motif based on the motif enrichment analysis.
Genes were identified in the motif enrichment analysis with pValue for the W-box motifs < = 0.001. Five modules were identified and labeled with the name of the representative gene. Red nodes – genes whose promoters contain the W-box motif; white nodes – genes whose promoters lack the W-box motif.
Table 3.
GO enrichment of co-expression modules identified in the W-box motif sub-network*.
Figure 7.
A sub-network for the W-box motif based on motif position bias analysis.
Genes were identified in the motif position analysis with z-score> = 3. Five modules were identified. Among them, 3 modules (circled in blue) were also identified via the motif enrichment analysis (See Figure 6), while 2 modules (circled in green) were additional modules identified via motif position analysis. Red nodes – genes whose promoters contain the W-box motif; white nodes – genes whose promoters lack the W-box motif.
Figure 8.
A sub-network for the site II element motif based on motif position bias analysis.
Genes were identified in the motif position analysis with z-score> = 3. Thirteen modules were identified, labeled with the name of the representative gene. Red nodes – genes whose promoters contain the site II element motif; white nodes – genes whose promoters lack the site II element motif.
Table 4.
GO enrichment of co-expression modules identified in the W-box motif sub-network*.
Figure 9.
A sub-network with the top 6,000 co-expressed gene pairs extracted from the whole gene co-expression network.
(A) Genes identified to be regulated by the G-box, MYB, W-box, and site II element motifs are spread across this sub-network, as indicated by the colors of the nodes. Nodes without colors are genes not identified to be regulated by these motifs. Circled are two modules that recapitulate the results from single motif analysis. (B) A Venn diagram showing the number of genes regulated by individual motifs or by combination of two motifs.
Table 5.
Comparison between the bottom-up approach (for GGM network) with the top-down approach (for AGCN network) on module discovery.
Figure 10.
A TF-promoter screening system based on the SGR gene.
(A) Transient overexpression of the SGR gene (left two leaves) in N. benthamiana induced yellowing, while a control gene did not (two leaves, upper right). (B) The design scheme for the screening system. (C) Transient over-expression of the SUR_Promoter::SGR construct together with 6 different MYBs (#1 to #6), and an actin gene as negative control (#0). Only MYB28 (#1), MYB29 (#2) induced yellowing.
Table 6.
TF-Promoter interaction verified by SGR- and luciferase-based screening.
Figure 11.
Gene expression modules regulated by transcription factors in a gene co-expression network.
(A) Different modules in a gene co-expression network are regulated by different transcription factors (TF-A, TF-B, TF-C etc). These transcription factors are not necessary part of the co-expression network, and they might interact with each other and form a regulatory network of themselves. (B) A gene expression model derived from the network analysis. The MYB transcription factors (MYB14, 15, and/or 32) are activated upon pathogens infection and turn on the expression of down-stream lignin biosynthesis genes (described in Figure 4 Module V).