Integrating images from multiple microscopy screens reveals diverse patterns of change in the subcellular localization of proteins.
Integrating images from multiple microscopy screens reveals diverse patterns of change in the subcellular localization of proteins.
复制标题
DOI：
10.7554/elife.31872
复制
发表时间：
2018-04-05
期刊：
影响因子：
7.7
通讯作者：
Moses AM
中科院分区：
文献类型：
--
作者：
Lu AX;Chong YT;Hsu IS;Strome B;Handfield LF;Kraus O;Andrews BJ;Moses AM
在系统水平上对蛋白质定位变化的评估是了解细胞如何响应环境、化学或遗传扰动的有力工具。到目前为止，通过高通量成像了解这些蛋白质组反应的工作已经独立地对每个扰动的定位变化进行了分类。为了区分这些变化是对特定扰动或更普遍的程序的有针对性的反应，我们开发了一种可扩展的方法，将多个实验中蛋白质的定位行为可视化为定量模式。通过将这种方法应用于由近400,000张图像组成的24个实验屏幕，我们区分了特定的反应和更普遍的反应，发现了应激反应蛋白定位行为的细微差别，并通过聚集具有相似模式的蛋白质形成了假设。以前的方法旨在尽可能准确地捕捉单个屏幕的所有定位变化，而我们的工作旨在整合大量成像数据，以发现意想不到的新细胞生物学。细胞是忙碌的结构，蛋白质在其中从一个地方移动到另一个地方以发挥他们的作用。为了生存，生物体必须不断适应变化，如突变、新环境或遇到药物。通常，这些适应性反应涉及一组蛋白质移动到细胞中的新位置。跟踪这些运动有助于理解细胞激活哪些生物过程来应对修饰。为了研究这些机制，科学家们进行了实验，让细胞暴露于一种类型的变化--例如，他们突变了一个基因，或者给了一种药物。然后，他们使用强大的显微镜跟踪细胞中的蛋白质。这些仪器每天可以拍摄数以千计的细胞照片，从而产生大量数据。然后对图像进行分析，这通常是主观的、耗时的，或者在其他实验中不容易重复。这阻止了研究人员同时考虑不止一个实验，这使得比较细胞对许多不同类型的变化的反应变得困难。在这里，Lu等人。他们结合和分析了对酵母菌进行的24次实验的40万张图像；他们使用一种计算方法，自动测量实验之间蛋白质位置的差异。这种新的大规模方法揭示了细胞生物学的一些方面，这些方面仅从任何一项实验的结果来看并不明显。例如，特定于一种类型变化的反应可以与每次细胞暴露在新条件下时发生的反应区分开来。这也是第一次有可能将在不同类型的变化中以相同方式移动的蛋白质组合在一起，并推断它们的作用。例如，一种蛋白质根据它与其他脉冲蛋白质共享某些特征的方式，被证明是“脉动的”--在细胞中的两个特定隔间之间快速移动。这种计算方法不仅限于对酵母的实验，也可以用于对人类细胞的研究。然后，制药公司可以大规模比较不同的治疗方法对蛋白质定位的影响。
The evaluation of protein localization changes on a systematic level is a powerful tool for understanding how cells respond to environmental, chemical, or genetic perturbations. To date, work in understanding these proteomic responses through high-throughput imaging has catalogued localization changes independently for each perturbation. To distinguish changes that are targeted responses to the specific perturbation or more generalized programs, we developed a scalable approach to visualize the localization behavior of proteins across multiple experiments as a quantitative pattern. By applying this approach to 24 experimental screens consisting of nearly 400,000 images, we differentiated specific responses from more generalized ones, discovered nuance in the localization behavior of stress-responsive proteins, and formed hypotheses by clustering proteins that have similar patterns. Previous approaches aim to capture all localization changes for a single screen as accurately as possible, whereas our work aims to integrate large amounts of imaging data to find unexpected new cell biology. Cells are busy structures in which proteins move from one place to the other to perform their roles. To survive, organisms must constantly adapt to changes, such as mutations, new environments, or encountering drugs. Often, these adaptive responses involve groups of proteins traveling to new locations in the cell. Tracking these movements is useful to understand which biological processes cells activate to cope with modifications. To study these mechanisms, scientists conduct experiments where they expose cells to one type of change – for example they mutate one gene, or give one drug. They then follow the proteins in the cell using powerful microscopes. These instruments can take thousands of pictures of cells every day, which results in large amounts of data. The images are then analyzed, which is often subjective, time-consuming or cannot easily be repeated on other experiments. This prevents researchers from considering more than one experiment at the time, and it makes it difficult to compare how cells respond across many different types of changes. Here, Lu et al. combine and analyze 400,000 images from 24 experiments conducted on yeasts; they use a computational method that automatically measures the differences in the locations of the proteins between experiments. This new large-scale approach reveals aspects of cell biology that are not obvious from the results of any one experiment alone. For example, responses that are specific to one type of change can be distinguished from the ones that occur each time cells are exposed to new conditions. It also becomes possible, for the first time, to group together proteins that move in the same way across different types of changes, and to infer their roles. For instance, a protein was shown to be ‘pulsing’ – moving quickly between two specific compartments in the cell – based on how it shared certain features with other pulsing proteins. This computational approach is not limited to experiments on yeasts, and could be used on studies in human cells as well. Drug companies could then compare at a large scale how different treatments affect protein localization.