Constructing and Analyzing Gene Regulatory Networks in Leaf Senescence of Arabidopsis thaliana

Abstract

Genes are known as the fundamental physical and functional building-blocks inherent in living organisms. Hence the detection of the behaviors of genes and how genes and their regulators interact with each other to carry out certain functions has fascinated researchers for decades. Recent advances in computational biology in modeling and constructing biological networks has leveraged the studies in molecular biology to arrive at new discoveries. Gene regulatory networks (GRNs) express the underlying gene-gene regulatory mechanisms in living organisms. Computational inference models have been introduced to construct these biological interaction networks in order to avoid the high cost and time consuming wet lab experiments in constructing GRNs. Only a limited number of computational models of GRNs currently exist, they too built for a few simple model organisms. This research concerns the constructing of GRNs for the process of leaf senescence in Arabidopsis thaliana as an example, to uncover the cell mechanisms that are responsible for aging. Even though leaf senescence of a plant is highly regulated and complex process, analysis of gene expression data and transcriptional factors enable connection of metabolic pathways and signaling pathways which then unpin development of gene regulatory network models to elucidate the process of leaf senescence. This study present a Bayesian network model to infer Gene regulatory networks for leaf senescence of Arabidopsis thaliana model plant. The networks were obtained con- sidering the gene expression values of up regulated and down regulated genes occurred in the process of leaf senescence. The obtained six Dynamic Bayesian networks which can be used to study the behavioral changes of leaves with time. Our dynamic GRN networks predict a number of valid gene-gene interactions which we believe directly im- pact leaf senescence of Arabidopsis. Up regulated and down regulated gene networks at di erent time periods of Arabidopsis leaf senescence were then analyzed and compared. Intuitively, we believed that the analysis of Gene network topologies could contribute to the understanding of the leaf senescence process. Several numbers of sensitive measures in network topologies were applied to the ob- tained dynamic senescence speci c GRNs. The gained knowledge has been used in under- standing the topological changes of networks with time and senescence. Results illustrate that there are signi cant changes in local topologies of networks than global properties highlighting the loss of connectivity among genes in the activation of visible senescence of a leaf. The Gene Ontology (GO) enrichment analysis reveals certain genes in signaling pathways which regulate in leaf senescence. This study successfully validated number of Senescence Associated genes (SAGs) and several gene-gene interaction predictions de- rived from the network inference phase.