An interactive tool for in-depth and reproducible analysis of RNA-seq data

项目总结 对于许多生物学家来说，大型基因组数据集的生物信息学分析是一个关键障碍，特别是 那些在较小的研究机构工作的人。利用我们团队的生物信息学经验，我们的目标是 开发可用于轻松翻译RNA测序数据的交互式Web应用程序 生物学的洞察力。我们假设，一个可重复的、深入分析的集成工具 表达数据将使高通量技术的使用大众化，并帮助生物学家精确定位 来自大数据的分子路径。我们的目标是开发一条精心设计的用户友好的管道 具有丰富的数据可视化能力。作为概念验证，该团队开发了一个名为IDEP的原型 (集成差异表达和通路分析)用于总结表达的分析 矩阵。它的独特功能包括(1)基于63 R和 BioConductor软件包，涵盖探索性数据分析、聚类、差异基因表达和 途径分析；(2)用于基因ID自动转换、注释和 2000多种古生物、细菌和真核生物的通径分析；(3)某些 通过生成R和R Markdown笔记本的核心步骤；(4)应用编程接口(API) 用于检索蛋白质-蛋白质相互作用网络和KEGG路径图，以及(5)易于访问 到9个物种约13000个经过处理的公共rna-seq数据。与现有工具相比，关键是 创新是强调深度集成(工具、注记、路径和公共数据集)、用户- 友好性和再现性。即使功能有限，IDEP也开始被 来自不同领域的研究人员。 在这项提案中，该团队计划完成IDEP的开发。具体目标1的目标是 要(A)以模块化、面向对象的方式重写IDEP，(B)制作一个R包，以便完全生成 可复制的R降价笔记本，以及(C)增加了必要的功能，如偏差纠正(批量 效果、GC含量、基因长度、表达水平)、时程分析、监督分类以及 现有功能模块的其他方法。我们还将实现基因本体的丰富 应用Blast2GO对未注解物种进行分析。具体目标2主要集中在(A) 扩大经常研究物种的途径数据库，并(B)更统一地收集 处理过的rna-seq和dna微阵列数据集，以促进对 公共表达数据。在具体目标3中，团队将进行硬件升级、严格测试、 代码审查、文档编制和社区集成。IDEP的发展可以帮助使 标准的rna-seq分析可供非常广泛的研究人员使用。

PROJECT SUMMARY Bioinformatic analysis of large genomic datasets is a critical barrier for many biologists, especially those at smaller research institutions. Leveraging our team's bioinformatics experience, our goal is to develop an interactive web application that can be used to easily translate RNA sequencing data into biological insights. We hypothesized that an integrated tool for reproducible, in-depth analysis of expression data will democratize access to high-throughput technologies and help biologists pinpoint molecular pathways from large data. Our goal is to develop a carefully-designed user-friendly pipeline with rich data visualization capacity. As a proof of concept, the team developed a prototype called iDEP (integrated Differential Expression and Pathway analysis) for the analysis of summarized expression matrices. It's unique features include (1) comprehensive analytic functionality based on 63 R and Bioconductor packages, covering exploratory data analysis, clustering, differential gene expression and pathway analysis; (2) a massive knowledgebase for automatic gene ID conversion, annotation, and pathway analysis for over 2000 archaeal, bacterial and eukaryotic species; (3) reproducibility of some core steps by generating R and R Markdown notebooks; (4) application programming interfaces (APIs) for retrieval of protein-protein interaction networks and KEGG pathway diagrams, and (5) easy access to about 13000 processed public RNA-seq data in 9 species. Compared with existing tools, the key innovation is the emphasis on deep integration (tools, annotation, pathways, and public datasets), user- friendliness, and reproducibility. Even with limited features, iDEP is beginning to be adopted by researchers from diverse fields. In this proposal, the team plans to complete the development of iDEP. The goal of Specific Aim 1 is to (a) re-write iDEP in a modular, object-oriented fashion, (b) make an R package for generating fully reproducible R Markdown notebooks, and (c) add essential functionalities such as bias correction (batch effect, GC content, gene length, expression level), time-course analysis, supervised classification, and additional methods for existing functional modules. We will also enable gene ontology enrichment analysis for unannotated species using Blast2GO. Specific Aim 2 focuses on (a) substantially expanding the pathway database for frequently studied species and (b) collecting more uniformly processed RNA-seq and DNA microarray datasets to facilitate the re-analysis and meta-analysis of public expression data. In Specific Aim 3, the team will conduct hardware upgrade, rigorous testing, code review, documentation, and community integration. The development of iDEP can help make standard RNA-seq analysis accessible for a very broad community of researchers.