AtGGM2014, an Arabidopsis gene co-expression network for functional studies

Abstract

Gene co-expression networks provide an important tool for systems biology studies. Using microarray data from the ArrayExpress database, we constructed an Arabidopsis gene co-expression network, termed AtGGM2014, based on the graphical Gaussian model, which contains 102,644 co-expression gene pairs among 18,068 genes. The network was grouped into 622 gene co-expression modules. These modules function in diverse house-keeping, cell cycle, development, hormone response, metabolism, and stress response pathways. We developed a tool to facilitate easy visualization of the expression patterns of these modules either in a tissue context or their regulation under different treatment conditions. The results indicate that at least six modules with tissue-specific expression pattern failed to record modular regulation under various stress conditions. This discrepancy could be best explained by the fact that experiments to study plant stress responses focused mainly on leaves and less on roots, and thus failed to recover specific regulation pattern in other tissues. Overall, the modular structures revealed by our network provide extensive information to generate testable hypotheses about diverse plant signaling pathways. AtGGM2014 offers a constructive tool for plant systems biology studies.Arabidopsis, gene co-expression network, graphical Gaussian model, plant development, stress response, hormone response Citation:Ma SS, Bohnert HJ, Dinesh-Kumar SP. AtGGM2014, an Arabidopsis gene co-expression network for functional studies .

Review of combinations of experimental and computational techniques to identify and understand genes involved in innate immunity and effector-triggered defence

“…In another study, high-scoring regulatory proteins (such as transcription factors, kinases and phosphatases) were fused into a co-expression network, enhancing correct prediction of phenotypes of known stress regulators from 36% to 62% [67]. Co-expression datasets of A. thaliana were used to show that biotic stress responses and hormone treatments induced several network modules and that hormones modified a module related to defence regulation, while effector proteins from P. syringae repressed two other modules [68]. Within these modules, genes of unknown function potentially contribute to plant defence responses.…”

Enhancing the biological relevance of Gene Co-expression Networks: A plant mitochondrial case study

“…One such approach, Gaussian Graphical Modelling, is commonly used as it allows researchers to interrogate the direct association between two genes, independent of the effects of surrounding genes present in the dataset. A number of thorough Gaussian Graphical Modelling studies in the model plant species Arabidopsis thaliana (Arabidopsis) have demonstrated the statistical power of this technique, and generated GCNs of select pathways and on a genome-wide scale (Wille et al, 2004; Ma et al, 2007; Ma et al, 2015). Another useful step for analysing complex datasets encompassing a wide range of tissues, developmental stages and stresses, is the use of batch-effect removal approaches.…”

Maize network analysis revealed gene modules involved in development, nutrients utilization, metabolism, and stress response

“…The original design was able to calculate Pcor data among a few thousand genes only [ 16 ]. Previously, we developed a random sampling-based method to overcome this obstacle and constructed the first genome-wide GGM network for Arabidopsis, followed by an updated network model termed AtGGM2014 [ 7 , 17 ]. Compared to other networks, AtGGM2014 contained more genes and identified additional modules participating in a large variety of plant processes, like development, metabolism, response to stresses, and response to hormones [ 7 ].…”

“…Previously, we developed a random sampling-based method to overcome this obstacle and constructed the first genome-wide GGM network for Arabidopsis, followed by an updated network model termed AtGGM2014 [ 7 , 17 ]. Compared to other networks, AtGGM2014 contained more genes and identified additional modules participating in a large variety of plant processes, like development, metabolism, response to stresses, and response to hormones [ 7 ]. For example, among many informative gene modules, it included hormonal signaling modules for phytohormones like auxin, abscisic acid, jasmonic acid, gibberellins, cytokinins, ethylene, and salicylic acid, demonstrating the network’s potential to facilitate systems biology studies on Arabidopsis gene functions.…”

“…Importantly, many of these modules contain transcription factor genes that act as potential gene expression regulator for the genes within the same modules. In addition, the maize network has been compared to a previous published Arabidopsis network [ 7 ]. This juxtaposition revealed conserved as well as diverging modules in the two species.…”