muscat detects subpopulation-specific state transitions from multi-sample multi-condition single-cell transcriptomics...

Area of Science:

• Genomics
• Computational Biology
• Bioinformatics

Background:

• Single-cell RNA sequencing (scRNA-seq) enables large-scale transcriptome profiling of individual cells.
• Current methods primarily focus on identifying subpopulation markers or cross-condition expression differences.
• Replicated multi-condition scRNA-seq datasets necessitate robust sample-level inference methods, termed differential state analysis.

Purpose of the Study:

• To survey and evaluate statistical frameworks for cross-condition differential state analysis in scRNA-seq data.
• To identify the most effective statistical approaches for handling replicated multi-condition scRNA-seq experiments.
• To provide reliable computational tools for multi-condition scRNA-seq data analysis.

Main Methods:

• Surveyed existing statistical methods, including cell-level mixed models and pseudobulk aggregated data approaches.
• Developed a flexible simulation framework to mimic multi-sample scRNA-seq data for performance evaluation.
• Applied methods to analyze scRNA-seq data from mouse cortex cells to assess responses to lipopolysaccharide treatment.

Main Results:

• Identified and compared the performance of various statistical methods for differential state analysis.
• Demonstrated the utility of these methods in uncovering subpopulation-specific biological responses.
• Validated findings using a simulated dataset that accurately reflects multi-sample scRNA-seq data characteristics.

Conclusions:

• Differential state analysis is crucial for interpreting complex, replicated scRNA-seq experiments.
• The study provides a comparative evaluation of statistical methods, guiding researchers in selecting appropriate frameworks.
• The developed tools within the muscat R package offer robust solutions for multi-condition scRNA-seq data analysis.