ARGLRR: A Sparse Low-Rank Representation Single-Cell RNA-Sequencing Data Clustering Method Combined with a New Graph...

Area of Science:

• Genomics
• Computational Biology
• Bioinformatics

Background:

• Single-cell RNA sequencing (scRNA-seq) enables cellular-level biological studies.
• Clustering scRNA-seq data identifies distinct cell types and reveals heterogeneity.
• scRNA-seq data's high noise and low coverage challenge existing clustering accuracy.

Purpose of the Study:

• To propose a novel sparse subspace clustering method for cell type identification in scRNA-seq data.
• To address the limitations of traditional Low-Rank Representation (LRR) methods in capturing local data structures.
• To improve the accuracy of cell type identification by integrating local and global data features.

Main Methods:

• Introduced Adjusted Random walk Graph regularization Sparse Low-Rank Representation (ARGLRR).
• Incorporated adjusted random walk graph regularization to capture local data structures.
• Applied similarity constraints within the LRR framework to enhance cell-to-cell similarity estimation.

Main Results:

• ARGLRR demonstrated superior performance on nine scRNA-seq datasets compared to advanced methods.
• Achieved 6.99% and 5.85% higher performance in Normalized Mutual Information and Adjusted Rand Index, respectively.
• Uniform Manifold Approximation and Projection visualization confirmed enhanced separation of cell types.

Conclusions:

• ARGLRR effectively captures both local and global structures in scRNA-seq data.
• The proposed method significantly improves cell type identification accuracy.
• ARGLRR offers a robust solution for analyzing complex scRNA-seq data and understanding cellular heterogeneity.