scEFSC: Accurate single-cell RNA-seq data analysis via ensemble consensus clustering based on multiple feature...

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Single-cell RNA sequencing (scRNA-seq) is crucial for understanding cellular heterogeneity.
• Clustering scRNA-seq data faces challenges like high dimensionality, numerical instability, and scalability.
• Existing clustering algorithms struggle with large cell numbers and high dropout rates.

Purpose of the Study:

• To systematically evaluate feature selection and clustering algorithms for scRNA-seq data.
• To propose a novel ensemble clustering method, scEFSC, addressing current limitations.
• To enhance the accuracy and biological interpretability of single-cell data clustering.

Main Methods:

• Evaluated four feature selection methods and nine clustering algorithms on fourteen scRNA-seq datasets.
• Developed scEFSC, employing unsupervised feature selection to filter non-significant genes.
• Combined clustering results from filtered data using weighted-based meta-clustering.

Main Results:

• scEFSC demonstrated superior performance across multiple evaluation metrics compared to existing algorithms on fourteen real scRNA-seq datasets.
• The method effectively handles high dimensionality and computational challenges inherent in scRNA-seq analysis.
• Biological interpretability was confirmed through differential gene expression, gene ontology, and KEGG pathway analyses.

Conclusions:

• scEFSC offers an accurate and robust approach for scRNA-seq data clustering.
• The ensemble feature selection strategy effectively improves clustering performance and scalability.
• scEFSC provides valuable biological insights, making it a promising tool for single-cell data analysis.