scDDboost

scDDboost analyzes changes in the distribution of single-cell expression data between two experimental conditions using an empirical Bayesian formulation that incorporates cell clustering information.

Key Features:

• Clustering-informed differential expression: Benefits uniquely from information conveyed by the clustering of cells into cellular subtypes.
• Empirical Bayesian posterior probabilities: Calculates gene-specific posterior probabilities that the marginal expression distribution is the same or different between two conditions.
• Mixture of negative binomial distributions: Treats gene-level expression data within each condition as a mixture of negative binomial distributions.

Scientific Applications:

• Single-cell RNA-seq differential expression analysis: Assesses changes in expression distributions between two experimental conditions from single-cell RNA-seq data.
• Cellular subtype-aware differential testing: Detects differential expression patterns that leverage underlying cell subtype structure in single-cell datasets.

Methodology:

scDDboost uses a novel empirical Bayesian formulation to calculate gene-specific posterior probabilities of distributional equivalence or difference between conditions, modeling gene-level expression data within each condition as a mixture of negative binomial distributions.