scde

scde models technical noise and expression-magnitude distortions in single-cell RNA sequencing (RNA-seq) data to enable robust differential expression analysis and identification of cellular subpopulations.

Key Features:

• Individual error models: Fits individual error models to single-cell RNA-seq measurements to account for cell- and gene-specific variability.
• Probabilistic expression-magnitude correction: Implements a probabilistic model specifically targeting expression-magnitude distortions observed in single-cell RNA-seq.
• Differential expression detection: Enhances detection of differential expression signatures across defined cell groups.
• Pagoda framework integration: Incorporates the pagoda framework to extend analyses beyond single genes to pathways and gene sets.
• Pathway and gene set overdispersion: Applies pathway- and gene set-level overdispersion analysis to detect coordinated transcriptional variation.
• Subpopulation identification: Facilitates identification and characterization of putative cell subpopulations based on transcriptional signatures.
• Noise and variability mitigation: Provides methods to mitigate technical noise and biological variability inherent to single-cell RNA-seq data.
• Single-cell expression exploration: Enables detailed exploration of gene expression patterns at the single-cell level.

Scientific Applications:

• Differential expression analysis: Detecting genes with differential expression across cell groups in single-cell RNA-seq experiments.
• Cellular subpopulation discovery: Identifying and characterizing subpopulations within complex tissues or specialized cellular environments.
• Pathway-level heterogeneity analysis: Revealing coordinated transcriptional programs and pathway overdispersion across cells.
• Tissue composition and diversity studies: Investigating tissue composition, cellular diversity, and mechanisms driving differential gene expression across cell populations.

Methodology:

Fits individual error models to single-cell RNA-seq measurements, implements a probabilistic model to account for expression-magnitude distortions, performs differential expression detection across cell groups, and applies the pagoda framework for pathway and gene set overdispersion analysis.