UICPC: Centrality-based clustering for scRNA-seq data analysis without user input.

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Single-cell RNA sequencing (scRNA-seq) analysis reveals novel cell types and heterogeneity.
• Existing scRNA-seq clustering methods often depend on user input, impacting performance.
• High dimensionality and heterogeneity in scRNA-seq data pose challenges for accurate cell grouping.

Purpose of the Study:

• To introduce UICPC, a novel centrality-clustering method for scRNA-seq data analysis.
• To evaluate UICPC's performance against state-of-the-art clustering techniques.
• To demonstrate UICPC's effectiveness in discovering cell groups without requiring user input.

Main Methods:

• Development of the UICPC centrality-clustering algorithm.
• Comparative analysis of UICPC with 9 existing clustering methods.
• Benchmarking on 11 diverse, real-world scRNA-seq datasets.

Main Results:

• UICPC demonstrates superior performance in cell group discovery compared to other methods.
• Key performance metrics including Normalized Mutual Information (NMI), Purity, and Adjusted Rand Index (ARI) show significant improvements with UICPC.
• The method shows robustness across datasets with high heterogeneity and large dimensions.

Conclusions:

• UICPC offers a user-input-free, effective solution for scRNA-seq data clustering.
• The method enhances the identification and characterization of cell populations.
• UICPC is available as an R package for broader research application.