作为在单个细胞分辨率水平研究转录组学的有效技术，单细胞RNA测序（scRNA-seq）近年来得到了快速发展并受到学术界日益重视，大大促进了复杂生物细胞异质性的研究和对疾病及其相关生物过程的理解。随著scRNA-seq实验和项目数量的迅猛增加，产生并积累了大量scRNA-seq数据，迫切需要有效的计算方法来处理这些数据。聚类分析是处理scRNA-seq数据的重要任务之一，既可以实现细胞分群、发现稀有细胞类型和推断细胞谱系，还可以探究细胞特异的生物特征与过程。本项目将从八个方面探索scRNA-seq数据聚类的新方法，包括：基于深度学习的聚类方法、小聚簇发现方法、基于稀疏表示的聚类方法、基于特征加权的聚类方法、基于主题模型的聚类方法、子空间聚类方法、基于细胞-基因图的聚类方法和快速聚类方法等。基于研究成果，将发表高水平学术论文，申请技术发明专利，实现开源共享的scRNA-seq数据聚类分析工具。

As an effective tool for interrogating the transcriptome at the resolution of individual cells, single-cell RNA-sequencing (scRNA-seq) has rapidly advanced and gained increasing attention from the academia in the past decade, which substantially pushes forward the research of intercellular heterogeneity and the understanding of diseases and associated biological processes. The increase of scRNA-seq experiments and projects leads to more and more scRNA-seq data. It is urgent to develop effective and efficient tools to handle such huge and increasing data. Clustering is an important task of analyzing scRNA-seq data, which can help us to group cells to different types, discover rare cell types, infer cell lineage, and explore cell-specific characteristics and processes of organisms. This proposal aims to develop effective methods for clustering scRNA-seq data in eight new technical perspectives, including clustering methods based on deep learning, topic models, feature weighting, sparse representation and cell-gene graphs, methods for discovering clusters of rare cells and fast clustering methods. This research is expected to produce high-quality publications, patents and open-source tools.