Nodulation Gene Networks in Legumes

With the accumulated -omics data sets, a major challenge has been to use these data to obtain a deeper understanding of the molecular mechanisms underlying complex traits. For example, root nodule symbiosis in legumes is one of the most productive nitrogen-fixing systems, fixing atmospheric nitrogen, and part of a sustainable agricultural cycle. Many genes and gene interactions involved in nodulation have been identified using traditional genetic and biochemical tools, but the complex nodule symbiosis process is far from fully understood. Here, three studies in gene network analysis have been done to help resolve this challenge in the post-genomics era. First, a nodulation gene network with 376 genes in Medicago truncatula was reconstructed using time-course transcriptome data during nodulation. Most of these genes are potentially novel nodulation-related genes. Their specific roles in nodulation have been predicted using module partition and functional analyses. Second, a new gene network reconstruction algorithm, weighted graphical lasso (wglasso), was developed to integrate multiple levels of -omics data. The algorithm significantly improved the accuracy of gene network reconstruction based on trancriptome data through the use other levels of gene interactions as prior knowledge. Third, a crowdsourcing platform was built for researchers to share and iteratively view nodulation gene network in Lotus japonicus. The comprehensive and accurate information for nodulation genes and gene interactions in the platform can be integrated into wglasso for better gene network reconstruction. Together, these products constitute a system that can be used to iteratively improve the prediction of nodulation gene networks. The system can be applied for other complex biological systems in order to quickly and systematically discover and understand molecular mechanisms underlying complex traits.