Collaborative Research: Characterizing Two Cell Polarity Processes Using Uncertainty Quantification to Analyze Complex Models and Data

该项目由数学科学部数学生物学计划和分子与细胞生物科学部细胞动力学和功能计划共同资助，其目标是更详细地了解细胞如何“生长”以产生不同的形状。具体来说，细胞如何重新排列自己，从一个对称的物体，如圆形的卵细胞，变成一个不对称的物体，如神经细胞？实验将在芽殖酵母上进行，芽殖酵母可以通过将组件定位到一端来从球形切换到不对称投影形状，这一过程称为细胞极化。极化的实例包括免疫细胞活化、肿瘤细胞转移和酵母感染，其中酵母细胞通过极化细胞生长侵入人体组织。真菌感染在美国的流行率正在增加，洞察力可以帮助研究人员发现阻止酵母感染侵入性生长的治疗方法。该研究将使用数学建模和显微成像相结合的方法来研究极化过程中细胞蛋白质的空间分布。实验数据可以与计算机模拟数据进行比较，以确认模型和调整模型参数。一个重要的方法上的进步将取代模型，这可能需要大量的时间来运行，与一个更简单的多项式代理模型，可以非常快速地计算，导致显着的计算速度。通过这一过程，人们获得了可以再现和预测极化细胞行为的模型。此外，研究还将与培养研究生、本科生和高中生如何进行定量显微镜实验和模拟数学模型的外展活动相结合。细胞极性和形态定义了个体和细胞群的形式和功能。系统生物学的方法将深入到这个问题在一个更定量的水平超越箭头图，以表征细胞极性的空间动态。本研究将利用芽殖酵母的实验易处理性来描述两种典型的细胞极性形态：芽和交配突起。更具体地说，这项研究将使用显微镜来可视化极化蛋白的空间动态，并将图像处理成定量数据。与此同时，一系列的数学模型的萌芽和交配投影生长将构建基于生物学假设。从方法论上讲，系统生物学的一个重大挑战是使用大型数据集估计模型/参数。贝叶斯推断将用于选择最佳模型并基于实验数据估计参数。该提案的一个核心概念是应用不确定性量化（UQ）技术，用替代多项式函数取代蒙特卡罗方法中的模型评估，从而大大加快不确定性分析的速度。结合的结果将是细胞极性的系统研究，导致模型预测，将通过实验测试转换成另一种细胞形态。此外，改进代理模型计算的拟议研究将进一步加速复杂模型的不确定性分析。该奖项反映了NSF的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The goal of this project, jointly funded by the Division of Mathematical Sciences Mathematical Biology Program and Division of Molecular and Cellular Biosciences Cellular Dynamics and Function Program, is to obtain a more detailed understanding of how cells "polarize" to generate different shapes. Specifically, how do cells rearrange themselves to go from a symmetric object like a round egg cell into an asymmetric object like a nerve cell? Experiments will be performed on budding yeast which can switch from a sphere to an asymmetric projection shape by localizing components to one end, a process termed cell polarization. Examples of polarization include immune cell activation, tumor cell metastasis, and yeast infections in which yeast cells invade human tissue by polarized cell growth. Fungal infections are increasing in prevalence in the U.S. Insights can help researchers discover treatments to halt the invasive growth of yeast infections. The research will investigate the spatial distribution of cellular proteins during polarization using a combination of mathematical modeling and microscopy imaging. Experimental data can be compared to computer simulation data to confirm the models and adjust the model parameters. An important methodological advance will be replacing the model, which can take significant time to run, with a simpler polynomial surrogate model which can be calculated very quickly resulting in a dramatic computational speed-up. Through this process one obtains models that can reproduce and predict the behavior of polarizing cells. In addition, the research will be integrated with outreach activities training graduate, undergraduate, and high school students on how to perform quantitative microscopy experiments and simulate mathematical models.Cell polarity and morphology define the form and function of individual and groups of cells. A systems biology approach will delve into this subject at a more quantitative level moving beyond arrow diagrams to characterize the spatial dynamics of cell polarity. The investigation will take advantage of the experimental tractability of budding yeast to characterize two classic cell polarity morphologies: the bud and the mating projection. More specifically, the investigation will use microscopy to visualize the spatial dynamics of polarization proteins and process the images into quantitative data. In parallel, a collection of mathematical models of budding and mating projection growth will be constructed based on biological hypotheses. Methodologically, one of the grand challenges of systems biology is to estimate the models/parameters using large datasets. Bayesian inference will be used to select the best models and estimate the parameters based on the experimental data. A central concept in this proposal will be applying techniques from uncertainty quantification (UQ) that replace model evaluations in the Monte Carlo method with a surrogate polynomial function, resulting in a dramatic speed-up of the uncertainty analysis. The combined result will be a systematic investigation of cell polarity leading to model predictions that will be tested by experiments converting one cellular morphology into the other. In addition, the proposed study of improved surrogate model calculation will further accelerate the uncertainty analysis of complex models.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.