Effective single-cell clustering through ensemble feature selection and similarity measurements

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Single-cell RNA sequencing (scRNA-seq) enables high-resolution gene expression profiling.
• Accurate cell type classification is crucial for scRNA-seq data analysis.
• Current clustering methods face challenges due to high dimensionality and dropout events.

Purpose of the Study:

• To develop an effective computational algorithm for single-cell clustering.
• To improve the accuracy and consistency of cell type identification in scRNA-seq data.
• To address the challenges of feature selection and noise in scRNA-seq analysis.

Main Methods:

• Ensemble feature selection to identify optimal genes.
• Cell-to-cell similarity measurement using multiple feature sampling.
• Construction of an ensemble network based on similarity.
• Application of a network-based clustering algorithm.

Main Results:

• The proposed algorithm achieves accurate and consistent single-cell clustering.
• Demonstrated effectiveness on real-world scRNA-seq datasets with known cell types.
• The method is compatible with existing analysis pipelines by accepting relative expression as input.

Conclusions:

• The developed algorithm provides a robust solution for single-cell clustering.
• Ensemble feature selection and similarity measurements enhance clustering performance.
• The publicly available source code facilitates adoption and further research.