dreamlet

dreamlet performs scalable differential expression analysis of large-scale single cell and single nucleus transcriptomics datasets using a pseudobulk approach based on precision-weighted linear mixed models to identify genes differentially expressed with traits across subjects for each cell cluster.

Key Features:

• Pseudobulk approach: Aggregates single cell data to perform differential expression analysis across individuals rather than individual cells.
• Precision-weighted linear mixed models: Fits a regression model for each gene and cell cluster while accounting for repeated measures study designs, high dimensional batch effects, and varying sequencing depth or observed cells per biosample.
• Computational scalability: Substantially faster and uses less memory than existing workflows, enabling analysis of cohort-scale datasets with millions of cells from hundreds of subjects.
• Complex statistical modeling: Supports complex study designs and controls the false positive rate.
• dream differential expression analysis: Includes differential expression analysis for repeated measures.

Scientific Applications:

• Cell type-specific disease biology: Identifies cell type specific gene expression differences associated with disease state, such as Alzheimer's disease cases versus controls in postmortem brain tissue.
• Genetic regulation studies: Analyzes cell type specific gene expression differences associated with genetic regulation in cohort-scale datasets.
• Stimulus-response analysis: Tests differential expression across individuals associated with stimulus or experimental conditions.
• Multi-factor variation quantification: Quantifies variation in gene expression attributable to individual, tissue, time point, or technical variables.

Methodology:

dreamlet applies a pseudobulk approach and fits a precision-weighted linear mixed model for each gene and cell cluster to test differential expression across individuals associated with a trait of interest.