speckle

speckle analyzes single-cell RNA sequencing (scRNA-seq) data to detect and test for differences in cell type proportions across experimental conditions.

Key Features:

• Analysis of cell type proportions: Functions tailored to quantify and test variations in cell type proportions across experimental conditions such as disease, treatment, or development.
• Propeller methodology: A robust statistical approach that leverages biological replication and accounts for sources of variability in scRNA-seq to detect significant differences in cell type proportions.
• Complex experimental designs: Support for analyses involving multiple biological replicates and other complex study designs.
• Performance and validation: Evaluated via simulations across varied scenarios and demonstrated strong performance in detecting significant shifts in cellular composition.

Scientific Applications:

• Cell composition studies: Identifying shifts in the relative abundance of specific cell types associated with disease, treatment, or developmental processes.
• Case studies: Applied to biological questions in human heart development, aging processes, and COVID-19 disease severity.

Methodology:

The core computational method is the propeller statistical approach, which leverages biological replication, accounts for variability inherent in scRNA-seq data, and has been evaluated using simulation-based assessments; the method is implemented in an R package.