The emergence of single-cell sequencing technology enables people to observe cells with unprecedented precision. However, it is difficult to capture the information on all cells and genes in one single-cell RNA sequencing (scRNA-seq) experiment. Single-cell data of a single modality cannot explain cell state and system changes in detail. The integrative analysis of single-cell data aims to address these two types of problems. Integrating multiple scRNA-seq data can collect complete cell types and provide a powerful boost for the construction of cell atlases. Integrating single-cell multimodal data can be used to study the causal relationship and gene regulation mechanism across modalities. The development and application of data integration methods helps fully explore the richness and relevance of single-cell data and discover meaningful biological changes. Based on this, this article reviews the basic principles, methods and applications of multiple scRNA-seq data integration and single-cell multimodal data integration. Moreover, the advantages and disadvantages of existing methods are discussed. Finally, the future development is prospected.
单细胞测序技术的出现使得人们能够以前所未有的精度观测细胞。然而，单次单细胞转录组测序（scRNA-seq）实验难以捕获所有细胞和基因的信息，单个模态的单细胞数据无法详细阐释细胞状态和系统变化，单细胞数据的整合分析旨在解决这两类问题。整合不同来源的scRNA-seq数据，可以收集完整的细胞类型，为构建细胞图谱提供强大助力；整合多个模态的单细胞数据，可以研究模态间因果关系和基因调控机制。数据整合方法的开发与应用帮助充分挖掘单细胞数据的丰富性和相关性，发现有意义的生物学变化。基于此，本文综述了多源scRNA-seq数据整合和单细胞多模态数据整合的基本原理、方法和应用，并讨论了现有方法的优势和不足，最后对未来的发展前景予以展望。.
Keywords: cell atlas; cell type; data integration; multi-modality; single-cell RNA sequencing.