An Abiotic-Stress Conditioned Gene Regulatory Network in Rice Predicted Using an Ensemble of Reverse-Engineering Solutions

Drought and other environmental stressors trigger several metabolic changes in plants, and a part of this change is reflected in the transcriptome. It is known that a synergistic coordination of the regulatory machinery achieves, maintains and reciprocates the desired metabolic state of the cell. However, the fraction of stress response regulators that have phenotypic or genetic evidence in rice is very small. Hence, there is a critical need for complementary approaches to identify the regulatory networks that manifest themselves during abiotic stress. We essayed gene responses in a multitude of conditions and captured information about the behavior of gene communities, and their regulatory dynamics in the context of abiotic stress. We aggregated results from 4 reverse-engineering tasks into a final consensus network arranged as a ranked list of inferred TF-gene links. Known DNA binding domains and other validated TF-gene links are aggregated towards the top of our rankings, which partially but independently validates our predictions. Overcoming the current limitations of rank aggregation methods, we show that combining results from complementary softwares for RNA-seq analysis is more robust than the use of a single method. We explored the network structure of genes that are most responsive to drought treatments in the RNA-seq of Inflorescence and flag leaf. Finally, we present an interactive database that, under a drought-specific module-regulatory network setting, will allow discovery of potential regulators, as well as the functional context of user-defined query genes.