scFseCluster: a feature selection-enhanced clustering for single-cell RNA-seq data.

Area of Science:

• Genomics
• Computational Biology
• Bioinformatics

Background:

• Single-cell RNA sequencing (scRNA-seq) reveals cellular heterogeneity and functional diversity beyond bulk RNA sequencing.
• Clustering is crucial for scRNA-seq data analysis, identifying cell types and states.
• Existing computational methods face challenges with generalization and high computational costs.

Purpose of the Study:

• To introduce scFseCluster, a novel computational framework for scRNA-seq clustering.
• To address the limitations of existing methods by improving generalization and reducing computational cost.

Main Methods:

• Developed scFseCluster, a framework integrating a metaheuristic algorithm (Feature Selection based on Quantum Squirrel Search Algorithm).
• The algorithm extracts optimal gene sets to enhance cell clustering performance.
• Validated through simulation experiments and comparative studies on benchmark scRNA-seq datasets.

Main Results:

• scFseCluster demonstrated high performance on eight benchmark scRNA-seq datasets.
• The framework significantly outperformed seven State-of-the-Art clustering algorithms.
• Feature selection on high-variable genes was shown to substantially improve clustering outcomes.

Conclusions:

• scFseCluster is a versatile and effective tool for scRNA-seq data clustering.
• The integration of advanced feature selection enhances the accuracy and efficiency of cell type identification.
• This framework offers a valuable advancement for single-cell data analysis in biological research.