Systematic identification of cell-wall related genes in Populus based on analysis of functional modules in co-expression network

Abstract

The identification of novel genes relevant to plant cell wall (PCW) biosynthesis in Populus is a highly important and challenging problem. We surveyed candidate Populus cell wall genes using a non-targeted approach. First, a genome-wide Populus gene co-expression network (PGCN) was constructed using microarray data available in the public domain. Module detection was then performed, followed by gene ontology (GO) enrichment analysis, to assign the functional category to these modules. Based on GO annotation, the modules involved in PCW biosynthesis were then selected and analyzed in detail to annotate the candidate PCW genes in these modules, including gene annotation, expression of genes in different tissues, and so on. We examined the overrepresented cis-regulatory elements (CREs) in the gene promoters to understand the possible transcriptionally co-regulated relationships among the genes within the functional modules of cell wall biosynthesis. PGCN contains 6,854 nodes (genes) with 324,238 edges. The topological properties of the network indicate scale-free and modular behavior. A total of 435 modules were identified; among which, 67 modules were identified by overrepresented GO terms. Six modules involved in cell wall biosynthesis were identified. Module 9 was mainly involved in cellular polysaccharide metabolic process in the primary cell wall, whereas Module 4 comprises genes involved in secondary cell wall biogenesis. In addition, we predicted and analyzed 10 putative CREs in the promoters of the genes in Module 4 and Module 9. The non-targeted approach of gene network analysis and the data presented here can help further identify and characterize cell wall related genes in Populus.

Figure 1. A flowchart of the computational analysis pipeline for discovery of plant cell wall genes.

Figure 2. Properties of the PGCN.

(A) Network densities at different PCC cutoff values. (B) Distribution of the node degree for PGCN. (C) The distribution of the clustering coefficient () in relation to the node degree for PGCN. (D) The distribution of the in relation to the node degree for random network.

Figure 3. Co-expression relationship between transcription factors (TFs) in Module 4 of PGCN.

A node represents a TF gene and an edge indicates a significant co-expression relationship between two genes. The gene name of the Arabidopsis homologous gene is used for each TF.