European Molecular Biology Laboratory, European Bioinformatics Institute (EMBL-EBI), Wellcome Genome Campus, Hinxton, United Kingdom.
VHsquared, Cambridge, United Kingdom.
J Allergy Clin Immunol. 2018 Apr;141(4):1411-1416. doi: 10.1016/j.jaci.2017.12.992. Epub 2018 Feb 21.
There is a wealth of biological pathway information available in the scientific literature, but it is spread across many thousands of publications. Alongside publications that contain definitive experimental discoveries are many others that have been dismissed as spurious, found to be irreproducible, or are contradicted by later results and consequently now considered controversial. Many descriptions and images of pathways are incomplete stylized representations that assume the reader is an expert and familiar with the established details of the process, which are consequently not fully explained. Pathway representations in publications frequently do not represent a complete, detailed, and unambiguous description of the molecules involved; their precise posttranslational state; or a full account of the molecular events they undergo while participating in a process. Although this might be sufficient to be interpreted by an expert reader, the lack of detail makes such pathways less useful and difficult to understand for anyone unfamiliar with the area and of limited use as the basis for computational models.
Reactome was established as a freely accessible knowledge base of human biological pathways. It is manually populated with interconnected molecular events that fully detail the molecular participants linked to published experimental data and background material by using a formal and open data structure that facilitates computational reuse. These data are accessible on a Web site in the form of pathway diagrams that have descriptive summaries and annotations and as downloadable data sets in several formats that can be reused with other computational tools. The entire database and all supporting software can be downloaded and reused under a Creative Commons license.
Pathways are authored by expert biologists who work with Reactome curators and editorial staff to represent the consensus in the field. Pathways are represented as interactive diagrams that include as much molecular detail as possible and are linked to literature citations that contain supporting experimental details. All newly created events undergo a peer-review process before they are added to the database and made available on the associated Web site. New content is added quarterly.
The 63rd release of Reactome in December 2017 contains 10,996 human proteins participating in 11,426 events in 2,179 pathways. In addition, analytic tools allow data set submission for the identification and visualization of pathway enrichment and representation of expression profiles as an overlay on Reactome pathways. Protein-protein and compound-protein interactions from several sources, including custom user data sets, can be added to extend pathways. Pathway diagrams and analytic result displays can be downloaded as editable images, human-readable reports, and files in several standard formats that are suitable for computational reuse. Reactome content is available programmatically through a REpresentational State Transfer (REST)-based content service and as a Neo4J graph database. Signaling pathways for IL-1 to IL-38 are hierarchically classified within the pathway "signaling by interleukins." The classification used is largely derived from Akdis et al.
The addition to Reactome of a complete set of the known human interleukins, their receptors, and established signaling pathways linked to annotations of relevant aspects of immune function provides a significant computationally accessible resource of information about this important family. This information can be extended easily as new discoveries become accepted as the consensus in the field. A key aim for the future is to increase coverage of gene expression changes induced by interleukin signaling.
科学文献中蕴含着丰富的生物学途径信息,但这些信息分散在成千上万的出版物中。除了包含明确的实验发现的出版物外,还有许多其他出版物被认为是虚假的,无法重复,或者与后来的结果相矛盾,因此被认为是有争议的。许多途径的描述和图像都是不完整的、风格化的表示形式,假设读者是专家,并且熟悉该过程的既定细节,因此这些细节没有得到充分解释。出版物中的途径表示形式通常不能完整、详细和明确地描述所涉及的分子;它们的确切翻译后状态;或者没有完整说明它们在参与过程中经历的分子事件。尽管这对于专家读者来说可能足以进行解释,但由于缺乏细节,这些途径对于不熟悉该领域的人来说,其可用性和理解程度较低,并且作为计算模型的基础也用处有限。
Reactome 是一个免费获取的人类生物途径知识库。它通过使用正式和开放的数据结构进行人工填充,该数据结构将分子事件相互连接,这些事件详细描述了与已发表的实验数据和背景材料相关联的分子参与者。这些数据以途径图的形式在网站上以描述性摘要和注释的形式提供,还可以以几种可重复使用的格式下载数据集,这些数据集可以与其他计算工具一起使用。整个数据库和所有支持的软件都可以根据知识共享许可协议下载和重复使用。
途径由具有生物学专业知识的专家生物学家编写,他们与 Reactome 的策展人和编辑人员合作,以代表该领域的共识。途径以交互式图表的形式呈现,其中包含尽可能多的分子细节,并链接到包含支持实验细节的文献引用。所有新创建的事件都要经过同行评审过程,然后才能添加到数据库中,并在相关网站上提供。每季度都会添加新内容。
2017 年 12 月发布的第 63 个 Reactome 版本包含 10996 个人类蛋白,参与 2179 条途径中的 11426 个事件。此外,分析工具允许提交数据集,以识别和可视化途径富集,并将表达谱表示为 Reactome 途径的叠加。来自几个来源的蛋白质-蛋白质和化合物-蛋白质相互作用,包括自定义用户数据集,可以添加到途径中以进行扩展。途径图和分析结果显示可以下载为可编辑图像、人类可读报告以及适合计算重用的几种标准格式的文件。Reactome 内容可通过基于表示状态转移 (REpresentational State Transfer,REST) 的内容服务和 Neo4J 图形数据库以编程方式获取。IL-1 至 IL-38 的信号通路在“白细胞介素信号”途径中进行了分层分类。所使用的分类主要来自 Akdis 等人的研究。
在 Reactome 中添加了一整套已知的人类白细胞介素、它们的受体以及与免疫功能相关方面的注释相关联的既定信号通路,为该重要家族提供了一个具有重要计算可访问性的信息资源。随着新发现被接受为该领域的共识,这些信息可以很容易地扩展。未来的一个关键目标是增加白细胞介素信号诱导的基因表达变化的覆盖范围。