cytoKernel: robust kernel embeddings for assessing differential expression of single-cell data.

Area of Science:

• Single-cell omics
• Computational biology
• Biostatistics

Background:

• High-throughput single-cell sequencing enables cell specification evaluation and identification of intricate variations.
• Existing differential expression methods often focus on aggregate measurements, missing subtle, multimodal expression changes.

Purpose of the Study:

• To introduce cytoKernel, a novel kernel-based score test for robust differential expression analysis in single-cell data.
• To develop a method capable of detecting both global and elusive differential expression patterns.

Main Methods:

• cytoKernel utilizes kernel embeddings to analyze the full probability distribution of single-cell RNA sequencing and cytometry data.
• It calculates pairwise divergence between subject distributions to identify differential expression patterns.

Main Results:

• cytoKernel effectively controls the false discovery rate and outperforms existing methods in benchmarks.
• The method successfully identifies more differential expression patterns, including subtle variations.
• Applied to real datasets, cytoKernel reveals gene and protein expression differences in cell subpopulations.

Conclusions:

• cytoKernel provides a powerful and sensitive approach for differential expression analysis in single-cell studies.
• The method enhances the ability to detect complex biological variations in high-dimensional single-cell data.