iDESC: identifying differential expression in single-cell RNA sequencing data with multiple subjects.

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Single-cell RNA sequencing (scRNA-seq) provides transcriptome-wide insights at single-cell resolution.
• Subject-specific variations and prevalent dropout events complicate differential expression (DE) analysis in scRNA-seq data.
• Accurate DE analysis is crucial for understanding cell type-specific responses in complex biological systems.

Purpose of the Study:

• To develop a novel method, iDESC, for robust cell type-specific differential expression analysis in multi-subject scRNA-seq data.
• To address the confounding effects of subject variability and dropout events in scRNA-seq data.
• To improve the accuracy and reliability of identifying differentially expressed genes.

Main Methods:

• Developed iDESC, employing a zero-inflated negative binomial mixed model.
• Modeled dropout event prevalence as a function of gene expression level.
• Incorporated subject effect as a random effect within the model framework.
• Evaluated performance against eleven existing DE analysis methods using simulated and real scRNA-seq datasets.

Main Results:

• iDESC demonstrated well-controlled Type I error rates and superior power in simulated data compared to existing methods.
• Application to three real scRNA-seq datasets revealed that iDESC results exhibited the highest consistency across datasets.
• iDESC analysis showed enhanced relevance to disease-specific variations.

Conclusions:

• iDESC effectively separates subject-specific effects from biological effects (e.g., disease) in scRNA-seq data.
• The method provides more accurate and robust differential expression analysis by explicitly modeling dropouts.
• Accounting for subject effects and dropouts is essential for reliable DE analysis in multi-subject scRNA-seq studies.