Dynamic regulatory network models of human response to influenza virus

项目总结 这个项目的目标是建立人类先天免疫反应的数学模型 病原体甲型流感病毒(IAV)。为了确保成功复制，病毒病原体必须同时劫持 宿主细胞机器的几个组件，同时逃避或使固有的细胞防御失效。这个 宿主遗传背景以及随后的病毒和宿主信号相互作用决定了疾病严重程度的范围 从无症状到死亡。最近在遗传多样性的小鼠模型中进行的IAV研究证实了关键的 基因型在宿主反应和结果中的作用。IAV的分子靶点和关键蛋白 发病机制可用于治疗以减轻或预防疾病。因此，我们构建了 驱动早期先天IAV反应的分子网络可以用来模拟遗传效应。我们的 实验系统是人肺上皮细胞，是IAV的第一道防线和靶点。 目的1.来自控制人上皮性IAV的基因调控网络(GRN)的遗传预测 回应。GRN描述转录因子(TF)对基因表达的控制。我们向大家展示了 将ATAC-SEQ与RNA-SEQ相结合可以提高GRN的精度。构建我们的动态GRN 异种肺组织模型中，我们提出了测量IAV感染的scRNA-seq和scATAC-seq以及 干扰素β刺激时程。我们小组最近发现了IAV蛋白NS1的新机制 驱动非依赖启动子的转录“通读”，并改变3D-染色质结构。因此，对于 模型中，我们还测量了基因组转录启动和启动子捕获Hi-C。后续实验 测试和GRN精化，我们将使用针对DNA序列和表观遗传学数据训练的深度学习模型 提供能够对数千种人类基因型态进行动态GRN模拟的输入。我们将确定 遗传风险基因座和驱动个体间基因表达反应差异的分子机制。 目的2.模拟蛋白质-蛋白质相互作用(PPI)和细胞信号网络驱动先天 对IAV的免疫反应我们开发了突变的流感病毒，每种病毒编码一种标记有旗帜的病毒蛋白， 同时保持体内的毒力。我们将使用突变的IAV来定位人类肺上皮细胞中的宿主病毒PPI 细胞和小鼠肺的活体实验。结合不同的组学数据，我们将构建一个分子网络 通过细胞信号通路将病毒-宿主PPI连接到IAV依赖的TF的模型。我们将测试 上位性标测的路径重建。 这两个目标的实现将导致GRN跨越病毒宿主PPI和细胞信号转导到TF控制 一种先天免疫细胞类型的基因表达。我们的实验-计算设计具有广泛的适用性。 这个模型，以及它未来对其他细胞的适应，将有助于识别遗传和分子机制 推动不同的人类IAV反应和可用于IAV治疗的网络漏洞。

PROJECT SUMMARY The goal of this project is to build mathematical models of human innate immune responses to the global pathogen influenza virus A (IAV). To ensure successful replication, viral pathogens must simultaneously hijack several components of the host cell machinery while either evading or disabling innate cellular defenses. The host genetic background and subsequent viral and host signaling interactions dictate disease severity ranging from asymptomatic to mortality. Recent studies of IAV in genetically diverse murine models confirm the critical role of genotype in host response and outcome. Both molecular targets as well as key proteins involved in IAV pathogenesis could be therapeutically exploited to attenuate or prevent disease. Thus, we construct models of the molecular networks driving early innate IAV response that can be used to model genetic effects. Our experimental system is human lung epithelium, the first-line of defense against and target of IAV. Aim 1. Genetic predictions from the gene regulatory network (GRN) governing human epithelial IAV response. GRNs describe the control of gene expression by transcription factors (TFs). We showed that integrating ATAC-seq with RNA-seq improves GRN accuracy. To construct a dynamic GRN in our heterogeneous lung tissue model, we propose scRNA-seq and scATAC-seq measurements of IAV infection and IFNβ stimulation time courses. Our group recently discovered new mechanisms by which the IAV protein Ns1 drives promoter-independent transcriptional “read-through” and alters 3D-chromatin architecture. Thus, for modeling, we also measure genomic transcription initiation and promoter-capture Hi-C. Following experimental testing and GRN refinement, we will use a deep-learning model trained on DNA sequence and epigenetic data to provide inputs that enable dynamic GRN simulations for thousands of human genotypes. We will identify genetic risk loci and molecular mechanisms driving difference in gene expression responses across individuals. Aim 2. Model the protein-protein interactions (PPIs) and cellular signaling networks driving the innate immune response to IAV. We developed mutant influenza viruses, each encoding a FLAG-tagged viral protein, while maintaining virulence in vivo. We will use the mutant IAV to map host-virus PPIs in human lung epithelial cells and mouse lung in vivo. Integrating with diverse ‘omics datasets, we will construct a molecular network model connecting virus-host PPIs through cellular signaling pathways to IAV-dependent TFs. We will test pathway reconstruction with epistasis mapping. Completion of both aims will lead to a GRN spanning virus-host PPIs and cellular signaling to TF control of gene expression in an innate-immune cell type. Our experimental-computational design is widely applicable. This model, and its future adaptation to other cells, will help identify the genetic and molecular mechanisms driving diverse human IAV responses and the network vulnerabilities to be exploited for IAV therapy.