Project 2

项目2 酵母菌落发育的多尺度时空动态 导言。 在生命系统中，个体的特征，包括对疾病的敏感性或对治疗的反应等特征，是由在不同组织规模上运作的过程的耦合决定的。例如，一个人的DNA序列限制了控制其细胞状态和行为的分子网络，这反过来又决定了多细胞结构的形式和功能。微生物，包括酵母，传统上被用作在单细胞水平上研究基本细胞过程的模型。然而，单细胞生物体可以形成多细胞群落，并分化成专门的结构，以造福于种群。在一些野生的酿酒酵母分离株中，菌落(从单个细胞开始，有丝分裂成为-10[8]细胞的结构)经历了形态转变，其特征是菌落表面有复杂的“皱纹”图案(图11)。 这一性状被称为“蓬松”表型。其他人的研究表明，蓬松的酵母菌落具有微生物生物膜的许多特性，因此与健康和人类疾病直接相关。在毛茸茸的菌落中，细胞由细胞外基质和内部中空通道连接，这可能有助于交换营养和废物。虽然有一些证据表明，这种形态发育涉及营养驱动的细胞状态转换、细胞-细胞信号、群体感应和细胞死亡，但确切的分子和在许多情况下的途径在很大程度上是未知的。 实验系统的重要性：这个项目试图了解细胞如何建立和维持细胞状态转变的时空模式，以形成多细胞结构。在我们的模型中(酵母“蓬松”的集落形成)，单个细胞分裂成10[8]个细胞的复杂结构，经历一系列的代谢和功能转变。通过提出下面的概念、计算和技术特征，我们将开发一种通用的方法来分析表现出形态表型的复杂特征，从而可以应用于各种问题，如发育期间的身体出生缺陷或肿瘤生长期间的血管生成。 挑战：？概念挑战：生物学中最基本的问题之一是基因型-表型问题：给定一个人的完整基因组序列，我们能预测这个人将表现出的特征吗？系统遗传学是一门新兴的学科，它通过将系统生物学的原理和技术应用于遗传分析来解决基因表型问题。一位少校 这种方法的局限性在于，虽然等式的基因型方是数据丰富的(例如，数十亿个核苷酸)，但特征是由一组极其有限的值(例如，血糖水平或发病年龄)定义的。 最近的方法，包括我们帮助开发的新的机器学习算法，已经试图通过使用诸如RNA表达水平的综合(组)数据来克服这种数据稀疏性问题。不幸的是，这些数据最能提供有关分子网络的信息(它们离分子网络更近)，并且从感兴趣的特征中删除了许多步骤，这可能是跨越多个生物尺度(分子网络、细胞、组织和器官)的过程的结果。推进系统遗传学在一个 为了减少这个问题的困惑，我们选择了一个有机体和一个特性，酵母中的菌落形态，其中成像的菌落结构和基因表达的空间模式可以用来提取一组与正在研究的特性直接相关的极其丰富的参数。

Project 2 Multiscale spatial and temporal dynamics of yeast colony development Introduction. In living systems, the characteristics of an individual, including traits such as susceptibility to disease or response to therapy, are determined by the coupling of processes that function at different scales of organization. For example, an individual's DNA sequence constrains the molecular networks that govern its cellular states and behaviors, which in turn determine the form and functions of multi-cellular structures. Microorganisms, including the yeast Saccharomyces cerevisiae, are traditionally used as models for investigating basic cellular processes at the unicellular level. However, unicellular organisms can form multi-cellular communities and differentiate into specialized structures to benefit the population. In some wild isolates of S. cerevisiae colonies (which start from a single cell and divide mitotically to become a structure of -10[8] cells) undergo a morphological transition characterized by complex patterns of "wrinkles" on the colony surface (Fig. 11). This trait is called the "fluffy" phenotype. Work by others has shown that fluffy yeast colonies possess many properties of microbial biofilms and are thus directly relevant to health and human disease. In fluffy colonies, cells are connected by an extracellular matrix and internal hollow channels, which may help exchange nutrients and waste products. While there is some evidence that this morphological development involves nutrient driven cell state transitions, cell-cell signaling, quorum sensing and cell death, the exact molecules and in many cases the pathways are largely unknown. The importance of the experimental system: This project seeks to understand how cells establish and maintain spatiotemporal patterns of cell state transitions to form multicellular structures. In our model (yeast "fluffy" colony formation) a single cell divides a undergoes a series of metabolic and functional transitions to reproducibly self organize into a complex structure of 10[8] cells. By advancing the conceptual, computational, and technical challeges below, we will develop a general methodology for analyzing complex traits that exhibit morphological phenotypes and can thus be applied to problems as diverse as physical birth defects during development or angiogenesis during tumor growth. The challenges: ¿ Conceptual challenges: Among the most fundamental problems in biology is the genotype-phenotype question: given the complete genome sequence of an individual, can we predict the traits that the individual will exhibit? The newly emerging field of systems genetics seeks to solve the genotype phenotype problem by applying the principles and technologies of systems biology to genetic analysis. A major limitation to this approach is that while the genotype side of the equation is data-rich (e.g., billions of nucleotides), traits are defined by an extremely limited set of values (e.g., a blood glucose level or age of disease onset). Recent methods, including new machine learning algorithms that we have helped develop, have attempted to overcome this data sparsity problem by using comprehensive ("omic") data, such as RNA expression levels. Unfortunately, such data are most informative about the molecular networks (to which they are more proximal) and many steps removed from the traits of interest, which can result from processes that span multiple biological scales (molecular networks, cells, tissues, and organs). To advance systems genetics in a way that is less confounded by this problem, we have chosen an organism and a trait, colony morphology in yeast, where imaging colony structure and spatial patterns of gene expression can be used to extract an extremely rich set of parameters that are directly relevant to the trait being studied.