An Effective Biclustering-Based Framework for Identifying Cell Subpopulations From scRNA-seq Data.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Single-cell RNA sequencing (scRNA-seq) enables transcriptomic analysis at the cellular level.
• Identifying distinct cell subpopulations is crucial for understanding cellular functions.
• High-dimensional, noisy, and sparse scRNA-seq data present significant analytical challenges.

Purpose of the Study:

• To introduce DivBiclust, a biclustering framework for robust cell subpopulation identification.
• To address the difficulties in clustering cells using high-dimensional and noisy scRNA-seq data.

Main Methods:

• Development of a biclustering-based framework named DivBiclust.
• Application and evaluation of DivBiclust on ten real-world scRNA-seq datasets.
• Comparison of DivBiclust against nine existing state-of-the-art methods.

Main Results:

• DivBiclust demonstrates superior accuracy in identifying cell subpopulations compared to existing methods.
• The framework effectively handles noisy and sparse scRNA-seq data.
• Experimental validation across diverse datasets confirms DivBiclust's performance.

Conclusions:

• DivBiclust provides an effective solution for cell subpopulation identification in scRNA-seq data analysis.
• The method offers high accuracy and robustness, even with challenging data characteristics.
• DivBiclust advances the analysis of single-cell transcriptomic data for functional cell studies.