BREM-SC: a bayesian random effects mixture model for joint clustering single cell multi-omics data.

Area of Science:

• Single-cell multi-omics analysis
• Computational biology
• Immunology

Background:

• Droplet-based single-cell RNA sequencing (scRNA-seq) enables large-scale gene expression profiling.
• Cellular Indexing of Transcriptomes and Epitopes by Sequencing (CITE-seq) integrates transcriptome and surface protein data.
• Existing computational tools struggle to analyze multi-modal CITE-seq data effectively.

Purpose of the Study:

• To develop a novel statistical method for joint analysis of single-cell transcriptomic and proteomic data.
• To address the lack of computational tools for multi-modal CITE-seq data analysis.
• To improve cell clustering accuracy by integrating both data types.

Main Methods:

• Developed BREM-SC, a Bayesian Random Effects Mixture model.
• Jointly clusters paired single-cell transcriptomic and proteomic data.
• Utilizes simulation studies and real-world datasets for validation.

Main Results:

• BREM-SC accurately identifies cell clusters by fully utilizing both transcriptomic and proteomic data.
• The method quantifies clustering uncertainty for individual cells.
• Demonstrated the validity and advantages of BREM-SC through comprehensive analyses.

Conclusions:

• BREM-SC offers a robust probabilistic approach for multi-modal single-cell data integration.
• Facilitates joint analysis of transcriptome and surface proteins at the single-cell level.
• Enables new biological discoveries, particularly in immunology research.