Collaborative Research: BBSRC: RiboViz for Reliable, Reproducible and Rigorous Quantification of Protein Synthesis from Ribosome Profiling Data

所有细胞都通过一种叫做核糖体的分子机器制造蛋白质，核糖体读取信使RNA模板并将RNA代码“翻译”成蛋白质代码。细胞需要在正确的时间以正确的数量制造正确的蛋白质，因此这个过程是由同样编码在RNA中的信号精心控制的。这些信号很复杂，只是刚刚开始被理解，因为细胞中有数千种不同的RNA序列，每个序列都有数百到数千个核苷酸（“字母”）长。DNA和RNA测序技术的最新进展意味着我们现在可以通过一种叫做核糖体分析的技术来测量被翻译成蛋白质的RNA的所有部分以及多少。尽管这项技术令人惊叹，但它并不完美，需要统计工具来将数据中有趣的生物信号与实验测量的不必要偏差分开。这些工具需要在可用且可靠的软件中实现，以便所有研究蛋白质合成的科学家能够从核糖体分析数据中获得最大可能的信息，而核糖体分析数据的收集既昂贵又耗时。RiboViz软件套件是开源的，世界上任何人都可以免费使用，它已经从测序机中获取原始数据，并将其经过一系列处理步骤。RiboViz估计RNA的每个部分被翻译了多少，以及翻译的数量是如何被RNA的代码控制的。RiboViz生成的表格、数字和图表都可以在线访问，因此对专家和非专家都很有用。这种数据共享使科学更容易再现，更容易获得。该项目将通过扩展RiboViz开源计算管道，从测量蛋白质合成的高通量数据中提取生物学见解，从而加速对蛋白质合成机制和调控的理解。目标是进一步开发RiboViz开源软件管道（https://github.com/shahpr/RiboViz），用于核糖体分析数据的可访问，可靠，可重复，严格和偏差感知的分析和可视化。具体目标是按照科学计算的最佳实践重构RiboViz，通过为设置、处理和分析的每个步骤编写类似于实验控制的测试，并通过管道的容器化，使其能够在不同的计算机上运行，并完全控制软件依赖关系；开发基于似然的统计方法，用于定量开放阅读框和密码子的差异翻译，同时纠正序列级偏差，建立差异RNA丰度分析的最佳实践，并在RiboViz中实施这些分析和可视化工具；并通过重新分析所有真核生物的已发表数据集来生成标准化的核糖体分析数据集，以严格量化密码子使用和其他序列特征如何预测蛋白质合成。改进后的RiboViz管道将加速翻译法规的研究，并产生经过测试和严格的工具，我们将作为开源资源传播给整个翻译研究社区。该奖项反映了美国国家科学基金会的法定使命，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。

All cells make proteins by using molecular machines called ribosomes, which read a messenger RNA template and "translate" the RNA code into the protein code. Cells need to make the right proteins, at the right time, in the right quantities, and so this process is carefully controlled by signals that are also encoded in the RNA. These signals are complex and only just beginning to be understood because there are thousands of different RNA sequences in a cell and each is hundreds to thousands of nucleotides ("letters") long. Recent advances in DNA & RNA sequencing technology mean that we can now measure all parts of RNA that are translated into protein and how much by using a technique called ribosome profiling. Although this technique is amazing, it is not perfect, and statistical tools are needed to separate the interesting biological signals in the data from unwanted biases of the experimental measurement. These tools need to be implemented in usable and reliable software in order for all scientists studying studying protein synthesis to be able to get the maximum possible information from ribosome profiling data, which is expensive and time-consuming to collect. The RiboViz software suite, which is open source and free to use by anyone in the world, already takes raw data from sequencing machines and puts it through a series of processing steps. RiboViz estimates how much each part of RNA is translated, and how the amount of translation is controlled by the code of that RNA. RiboViz produces tables, figures and graphs that are accessible online, so is useful for both experts and non-experts. This kind of data sharing makes science more reproducible and more accessible. This project will accelerate understanding of the mechanism and regulation of protein synthesis by extending the RiboViz open-source computational pipeline to extract biological insight from high-throughput data measuring protein synthesis. The goal is to further develop the RiboViz open-source software pipeline (https://github.com/shahpr/RiboViz) for accessible, reliable, reproducible, rigorous and bias-aware analysis and visualization of ribosome profiling data. Specific aims are to refactor RiboViz following best practices for scientific computing, by writing tests akin to experimental controls for each step of the setup, processing and analysis, and by containerization of the pipeline to enable running on different computers with full control of software dependencies; develop likelihood-based statistical methods for quantification of differential translation of open reading frames and codons while correcting for sequence-level bias, building on best practices in differential RNA abundance analysis, and implement these analysis and visualization tools within RiboViz; and to generate standardized ribosome profiling datasets by re-analyzing published datasets for all eukaryotes to quantify rigorously how codon usage and other sequence features predict protein synthesis. The improved RiboViz pipeline will accelerate studies of translation regulation and produce tested and rigorous tools that we will be disseminated as an open-source resource to the entire community studying translation.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.