switchde

switchde detects switch-like differential gene expression along pseudotemporal trajectories derived from single-cell RNA-seq (scRNA-seq) data.

Key Features:

• Fast Model Fitting: Employs efficient algorithms for rapid model fitting and returns interpretable parameter estimates that describe switch location and the rate of expression change along pseudotime.
• Statistical Significance Testing: Computes P-values to test switch-like models against a constant-expression null model.
• Zero-Inflation Modeling: Optionally incorporates zero-inflation models to account for excess zeros typical of scRNA-seq data.
• Implementation: Implemented as an R package associated with the Bioconductor project.

Scientific Applications:

• Cellular Differentiation: Identifies genes that undergo abrupt regulatory changes during lineage commitment and differentiation processes.
• Developmental Biology: Detects switch-like transcriptional events that mark developmental transitions along pseudotemporal trajectories.
• Cell Cycle and Temporal Processes: Reveals genes with stepwise expression changes associated with cell cycle progression and other dynamic cellular programs.

Methodology:

Constructs a statistical model capturing non-linear, stepwise dynamics of gene expression along pseudotime; fits model parameters with efficient algorithms to estimate switch location and rate, computes P-values against a constant-expression model, and optionally fits zero-inflation components for scRNA-seq data.