BAMM-SC

BAMM-SC performs Bayesian mixture modeling to cluster droplet-based single-cell RNA sequencing (scRNA-seq) raw count data from population-scale studies while accounting for inter-individual heterogeneity and batch effects.

Key Features:

• Bayesian Hierarchical Framework: Employs a unified Bayesian hierarchical mixture model that directly integrates raw scRNA-seq count data.
• Population-Scale Analysis: Scales to tens of thousands of single cells collected from multiple subjects to support cohort-level analyses.
• Inter-individual Heterogeneity and Batch Effect Handling: Explicitly models biological variability among cells and technical variability across individuals to mitigate batch effects.
• Validated Clustering Accuracy: Demonstrates improved clustering accuracy in simulation studies and experimental datasets including blood, lung, and skin cells from human and mouse.

Scientific Applications:

• Comparative Immunology: Analyzing immune cell populations across individuals to assess variation in immune responses.
• Developmental Biology: Identifying tissue-specific cell types and states across developmental stages or conditions.
• Disease Research: Characterizing cellular heterogeneity in disease states to inform pathogenesis and potential therapeutic targets.

Methodology:

Accepts raw count matrices from droplet-based scRNA-seq and models the data using a Bayesian mixture/hierarchical framework that accounts for biological variability among cells and technical variability across individuals.