UICPC: Centrality-based clustering for scRNA-seq data analysis without user input. (October 2021)

• Record Type: Journal Article
• Authors: Chowdhury, Hussain Ahmed; Bhattacharyya, Dhruba Kumar; Kalita, Jugal Kumar
• Is Part Of: Computers in biology and medicine. Volume 137(2021)
• Publication Date: 2021-10
• Subjects: Clustering -- scRNA-seq -- NGS -- Bioinformatics -- Computational biology -- Machine learning; Medicine -- Data processing -- Periodicals; Biology -- Data processing -- Periodicals; 610.285
• DOI: 10.1016/j.compbiomed.2021.104820
• Languages: English
• ISSNs: 0010-4825

Abstract

Abstract: scRNA-seq data analysis enables new possibilities for identification of novel cells, specific characterization of known cells and study of cell heterogeneity. The performance of most clustering methods especially developed for scRNA-seq is greatly influenced by user input. We propose a centrality-clustering method named UICPC and compare its performance with 9 state-of-the-art clustering methods on 11 real-world scRNA-seq datasets to demonstrate its effectiveness and usefulness in discovering cell groups. Our method does not require user input. However, it requires settings of threshold, which are benchmarked after performing extensive experiments. We observe that most compared approaches show poor performance due to high heterogeneity and large dataset dimensions. However, UICPC shows excellent performance in terms of NMI, Purity and ARI, respectively. UICPC is available as an R package and can be downloaded by clicking the link https://sites.google.com/view/hussinchowdhury/software . Graphical abstract: Image 1 Highlights: This paper proposes a user input free clustering method for scRNA-seq data. The method does not depend on any specific cluster validity index. Results show its superiority in terms of discovering true cell groups. Datasets and source code are available as an R package.