Integrating Multi-Scale Imaging, Reaction-Diffusion Simulation, and Markov Model Inference to Enhance Predictive Design and Interpretation of Single-Molecule Gene Regulation Experiments

项目摘要 单细胞成像可以量化基因调控的错综复杂的时空动态 从细菌感染到癌症的重要生物医学过程。这一基因调控受到 生物过程的复杂性和随机性，其观测还有待测量 由于生化标记的低效和显微镜成像的失真而产生的伪影。然而，尽管如此， 复杂性，初步工作表明，有可能整合数据和计算模型来预测 在各种环境和遗传条件下的基因调控条件：(1)模型必须受到约束 通过提供信息和可重现的数据，(2)模型必须经过严格验证，以解释生物学和 技术差异；(3)必须系统地探索模型，以量化不确定性。最重要的是 这个项目的假设是，在亚细胞动力学中观察到的时空波动包含 独特的信息，可以通过改进的计算方法和模型指导的实验来解锁。 为了验证这一假设，该项目将创建一个新的研究平台，被称为单细胞图形 解释和设计实验的实用工具。ScGUIDE将结合实验分析(例如图像处理 和单粒子跟踪以从荧光显微镜实验中提取定量数据)、空间 随机模拟(例如，生成模拟细胞实验的逼真视频的反应-扩散模型)， 模型抽象和识别(例如，要翻译的参数推理和不确定性量化 对预测性洞察的定量观察)和实验设计(例如，精确定位的统计方法 哪些具体的实验条件最有可能揭示新的生物学见解)。 为了展示其广泛的功能，将使用scGUIDE来分析和设计单电池 对四种不同的与健康相关的过程进行实验。在酵母中，该项目将检查协调 应激激活的MAPK动力学与SPT-Ada-Gcn5乙酰基转移酶(SAGA)亚单位之间的关系 染色质和RNA转录/运输动力学，它们与癌症、骨骼 发育不良和视网膜变性。在人类细胞中，该项目将研究时空聚集和 CDK抑制剂癌中RNAP聚合酶II参与单基因转录时的磷酸化 治疗。在骨肉瘤细胞中，该项目将探索对当地tRNA资源的竞争如何影响 单个信使核糖核酸分子在不同的亚细胞区域以及在人类和病毒环境中的翻译。最后， 该项目将探索表观遗传记忆和分子竞争对多基因的影响。 纸质操纵子的世代激活或抑制，使大肠杆菌能够建立尿路病原性感染。 每个项目都将建立空间和时间相互作用的机械性和定量预测模型 转录或翻译因子、酶和复杂分子机器与环境相结合 在单基因、单信使核糖核酸、单细胞和群体水平上调节表达的影响。

Project Summary Single-cell imaging can quantify intricate spatial and temporal dynamics of gene regulation that underly important biomedical process ranging from bacterial infections to cancer. This gene regulation is subject to complexities and randomness of biological processes, and its observation is further subject to measurement artifacts due to inefficiencies in biochemical labels and distortions in microscope imaging. Yet, despite these complications, preliminary work shows that it is possible to integrate data and computational models to predict gene regulation in myriad environmental and genetic conditions provided that: (1) models must be constrained by informative and reproducible data, (2) models must be rigorously verified to account for biological and technical variations, and (3) models must be systematically explored to quantify uncertainties. The overarching hypothesis of this project is that spatial and temporal fluctuations observed in subcellular dynamics contain unique information that can be unlocked with improved computational methods and model-guided experiments. To test this hypothesis, this project will create a new research platform to be known as the single-cell Graphical Utility to Interpret and Design Experiments. scGUIDE will combine experimental analysis (e.g., image processing and single-particle tracking to extract quantitative data from fluorescence microscopy experiments), spatial stochastic simulation (e.g., reaction-diffusion models to generate realistic videos to mimic cellular experiments), model abstraction and identification (e.g., parameter inference and uncertainty quantification to translate quantitative observations into predictive insight), and experiment design (e.g., statistical methods to pinpoint which specific experimental conditions are most likely to reveal new biological insight). To demonstrate its broad capabilities, scGUIDE will be used to analyze and design single-cell experiments for four different health-related processes. In yeast, the project will examine the coordination between stress-activated MAPK dynamics and Spt-Ada-Gcn5 Acetyltransferase (SAGA) subunits that control chromatin and RNA transcription/transport dynamics, and which have been implicated in carcinoma, skeletal dysplasia, and retinal degeneration. In human cells, the project will examine the spatiotemporal clustering and phosphorylation of RNAP Polymerase II as it engages in single-gene transcription under CDK-inhibitor cancer treatments. In osteosarcoma cells, the project will explore how competition for local tRNA resources affects translation of single-mRNA molecules in different sub-cellular regions and in human and viral contexts. Finally, the project will explore the effects that epigenetic memory and molecular competition have on the multi- generational activation or repression of the pap operon that allows E. coli to establish uropathogenic infections. Each project will build mechanistic and quantitatively predictive models for how spatial and temporal interactions of transcription or translation factors, enzymes, and complex molecular machines combine with environmental influences to regulate expression at the single-gene, single-mRNA, single-cell, and population levels.