The growing number of multi-omics studies demands clear conceptual workflows coupled with easy-to-use software tools to facilitate data analysis and interpretation. This protocol covers three key components involved in multi-omics analysis, including single-omics data analysis, knowledge-driven integration using biological networks and data-driven integration through joint dimensionality reduction. Using the dataset from a recent multi-omics study of human pancreatic islet tissue and plasma samples, the first section introduces how to perform transcriptomics/proteomics data analysis using ExpressAnalyst and lipidomics data analysis using MetaboAnalyst. On the basis of significant features detected in these workflows, the second section demonstrates how to perform knowledge-driven integration using OmicsNet. The last section illustrates how to perform data-driven integration from the normalized omics data and metadata using OmicsAnalyst. The complete protocol can be executed in ~2 h. Compared with other available options for multi-omics integration, the Analyst software suite described in this protocol enables researchers to perform a wide range of omics data analysis tasks via a user-friendly web interface.

越来越多的多组学研究需要清晰的概念工作流程以及易于使用的软件工具，以促进数据分析和解释。该协议涵盖了多组学分析中涉及的三个关键组成部分，包括单组学数据分析、使用生物网络的知识驱动集成以及通过联合降维的数据驱动集成。使用最近对人类胰岛组织和血浆样本进行的多组学研究的数据集，第一部分介绍如何使用 ExpressAnalyst 进行转录组学/蛋白质组学数据分析，以及如何使用 MetaboAnalyst 进行脂质组学数据分析。基于这些工作流程中检测到的重要特征，第二部分演示了如何使用 OmicsNet 执行知识驱动的集成。最后一部分说明如何使用 OmicsAnalyst 从标准化组学数据和元数据执行数据驱动的集成。完整的协议可以在约 2 小时内执行。与多组学集成的其他可用选项相比，本协议中描述的 Analyst 软件包使研究人员能够通过用户友好的 Web 界面执行各种组学数据分析任务。