A hybrid neighborhood enhanced contrastive learning and self-knowledge distillation method for scRNA-seq data...

Area of Science:

• Computational Biology
• Genomics
• Bioinformatics

Background:

• Single-cell RNA sequencing (scRNA-seq) data presents high dimensionality, complexity, and noise, challenging precise cell type classification.
• Existing analytical methods struggle with generalization and adaptability, leading to biased subpopulation identification and hindering biological understanding.

Purpose of the Study:

• To develop a novel method, scKD, for enhanced single-cell heterogeneity analysis and precise cell type classification.
• To improve clustering accuracy, stability, and robustness in scRNA-seq data analysis.

Main Methods:

• scKD integrates a hybrid neighborhood-enhanced comparative learning model.
• A self-knowledge distillation strategy is employed within the scKD framework.

Main Results:

• scKD demonstrates superior performance in identifying major cell types and rare cell subtypes.
• Extensive evaluations on multiple real-world datasets confirm scKD's robustness and clustering stability.
• The method achieves enhanced subpopulation identification accuracy.

Conclusions:

• scKD is a powerful and reliable tool for analyzing single-cell transcriptomic data.
• The proposed method facilitates deeper insights into cellular heterogeneity and biological processes.