Applications of Probabilistic Graphical Models in Genomic Networks for Agriculture
Abstract
Agricultural productivity is severely limited by environmental stresses that affect plants. Environmental stresses can be classified as abiotic or biotic. This study focuses on drought and saline stress, the two significant abiotic stresses causing crop loss worldwide. Crop loss due to drought and saline stress are major factors that threaten global food security. This problem is exacerbated by the growing world population, which is expected to rise by 2 billion in the next thirty years. Fortunately, plants have internal mechanisms to defend against environmental stresses. These mechanisms are deployed through complex networks of molecules known as signaling pathways. Environmental stress stimuli can trigger signaling pathways that activate or inhibit downstream genes to implement defensive measures and restore homeostasis. Signaling pathways are not only limited in their capability to defend against stresses but are also responsible for mediating other activities, including protein synthesis, cell death, and differentiation. Thus understanding the signaling pathways in plants is key to developing plants that can defend against environmental stresses and are nutritionally valuable.
We studied the drought signaling pathways in Arabidopsis to identify the genetic regulators
of drought-responsive genes. Additionally, we examined the lysine biosynthesis pathway in rice
under normal and saline stress conditions. Lysine is an essential amino acid present in the lowest quantity compared to all the other amino acids in rice. Amino acids are the building block of proteins and play a crucial role in maintaining the human body’s healthy functioning. Thus, increasing the lysine content in rice will help improve global health. We modeled both the drought signaling and lysine synthesis pathways using Bayesian networks. We chose Bayesian networks as they allow us to integrate pathway information from literature with experimental data. Using Bayesian networks, we identified that ATAF1 is a negative regulator of drought and DAPF is the most potent regulator of lysine. These regulators can be targeted using genetic intervention methods such as CRISPR-CAS9 to make plants robust against drought and increase lysine content in rice. Our work with drought signaling pathways was validated through wet-lab experiments.
Subject
DroughtLysine
Rice
Arabidopsis
Pathways
Bayesian Networks
Parameter Estimation
Network Modeling
Citation
Lahiri, Aditya (2022). Applications of Probabilistic Graphical Models in Genomic Networks for Agriculture. Doctoral dissertation, Texas A&M University. Available electronically from https : / /hdl .handle .net /1969 .1 /197202.