Factors Driving Diversity in Gene Regulatory Networks at Genome Scale[ASE]

Gene expression is a quantitative trait under the control of genetic and environmental factors and their interaction, so-called GxE. Understanding the mechanisms driving GxE is fundamental for ensuring stable crop performance across environments, and for predicting the response of natural populations to climate change. Gene expression is regulated through complex molecular networks, however environmental and genotypic effects on genome-wide regulatory networks are rarely considered. In this study, we model genome-scale gene expression variation between two natural accessions of the model grass Brachypodium distachyon and their response to soil drying. We identified genotypic, environmental, and GxE responses in physiological, metabolic, and gene expression traits. We then identified gene regulation conservation and variation among conditions and genotypes, simplified as co-expression clusters found unique in or conserved across library types. Putative gene regulatory interactions are inferred as network edges with a graphical model approach, resulting in hypotheses about gene-gene interactions which are then found to be specific to or with higher affinity in one genotype (G regulation), one environmental treatment (E regulation), or in one genotype under treatment (GxE regulation). Some gene-gene interactions are conserved across conditions so the differential expression is accordingly transmitted to target genes. These variably detected edges cluster together in co-expression modules, suggestive of different constraints or selection strength acting on specific pathways. We further applied our graphical model approach to identify putative, E-dependent regulatory mechanisms of leaf glucose content as an exemplar metabolite. Our study highlights an approach to identify variable features of gene regulatory networks and thereby identify key components for later genomic intervention to elucidate function or modulate environmental response. Our results also suggest possible targets of evolutionary change in gene regulatory networks associated with environmental plasticity. To study regulation mechanism of how different genotypes of Brachypodium (Bd21 and Bd3-1) response to soil water deficit differently, we generated hybrid of the two genotypes and applied same treatments and harvest samples for RNA sequencing. We started gradual dry-down on 33rd day with a soil water content ended up 55% for drought and 85% for control after 6 days. The youngest fully expanded leaves from 5 replicates of hybrids under drought and control treatment, and 2 control Bd3-1 samples were harvested at a single time point 1-2:30pm of the day for RNA sequencing. These along with 12 samples (each genotype and condition, replicated three times) from the main experiment with exact same treatments are sequenced. We did gene expression profiling analysis and differential expression by fitting models using DESeq2 to infer gene regulation mechanism. Grant: IOS 2239070 Grant title: NSF CAREER Award Grantee: David Des Marais Funding agency: National Science Foundation

基因表达是一类受遗传因素、环境因素及其互作（即G×E）调控的数量性状。解析G×E的作用机制，是保障作物在多样环境中稳定表现、预测自然种群对气候变化响应的核心基础。基因表达通过复杂的分子网络实现调控，但现有研究极少兼顾环境与基因型对全基因组调控网络的综合影响。本研究针对模式草本植物二穗短柄草（*Brachypodium distachyon*）的两个自然种质系及其对土壤干旱的响应，构建了全基因组范围的基因表达变异模型。我们在生理、代谢及基因表达三类性状中，分别鉴定出了基因型效应、环境效应及G×E互作效应。

随后，我们鉴定了不同条件与基因型间的基因调控保守性与变异，并将其简化为在特定文库类型中特有或跨文库类型保守的共表达簇。我们通过图形模型方法将潜在基因调控互作推断为网络边，进而得到基因-基因互作假说；这些互作被发现要么仅特异性存在于某一基因型（基因型调控）、某一环境处理（环境调控），要么仅在某一基因型接受处理时存在，或在该条件下互作强度更高（G×E调控）。部分基因-基因互作在不同条件下保守，因此差异表达信号可传递至下游靶基因。这些被可变检测到的网络边聚集于共表达模块中，提示特定通路受到不同的约束条件或选择压力作用。

我们进一步将该图形模型方法应用于叶片葡萄糖含量这一典型代谢物，鉴定了其依赖于环境的潜在调控机制。本研究提出了一种可鉴定基因调控网络可变特征的方法，可为后续通过基因组干预解析基因功能或调控环境响应的关键组分提供理论依据。本研究结果同时提示，与环境可塑性相关的基因调控网络存在进化改变的潜在靶点。

为解析二穗短柄草不同基因型（Bd21与Bd3-1）对土壤水分亏缺的差异化响应调控机制，我们构建了两个基因型的杂交群体，对其施加统一处理并采集样本用于RNA测序。实验于种植后第33天启动逐步干旱处理，持续6天后，干旱组土壤含水量最终维持在55%，对照组维持在85%。于处理完成当日的13:00-14:30这一时间点，采集干旱与对照处理下杂交群体的5份生物学重复样本的最新完全展开叶，同时采集2份Bd3-1对照样本用于RNA测序。此外，本研究主实验中12份样本（每个基因型与处理组合均设置3次生物学重复，处理条件完全一致）也一并完成了测序。我们通过DESeq2拟合模型开展基因表达谱分析与差异表达分析，以推断基因调控机制。

本研究受美国国家科学基金会（National Science Foundation, NSF）IOS 2239070号资助项目（NSF CAREER Award）资助，受资助人为David Des Marais。