Controllability Analysis of A Gene Network for Arabidopsis thaliana Reveals Characteristics of Functional Gene Families.

Based on structural controllability of complex networks and a constructed gene network with 9241 nodes for Arabidopsis thaliana, we classified nodes into five categories via their roles in control or node deletion, including indispensable, neutral, dispensable, driver and critical driver nodes. The indispensable nodes can increase the number of drivers after deletion, which are never drivers or critical drivers. About 10% nodes are indispensable. However, more than 60% nodes are neutral ones. More than 62% nodes are drivers, and indicates the gene network is very difficult to be fully controlled. Gene Ontology (GO) enrichment analysis reveals that different sets of nodes have preferred biological functions and processes.The indispensable nodes are significantly enriched as essential genes, drought responsive and abscisic acid (ABA) independent genes, transcriptional factors (TFs), core cell cycle genes, ABA and Gibberellin (GA) related genes. The critical drivers are enriched as receptor kinase-like genes, while shorted in WRKY TFs and functional genes that are enriched in the indispensable nodes. Robustness analysis based on node and edge additions, edge rewiring indicate the obtained conclusions are robust to network perturbations. Our investigations clarify control roles of some gene families and provide potential implications for identifying functional genes in other plant species, such as drought responsive genes and TFs.