M3S

M3S selects statistical models to capture multimodal gene expression distributions in single-cell RNA sequencing (scRNA-seq) data for accurate parameter estimation and differential expression analysis.

Key Features:

• Gene-Wise Model Selection: M3S selects the most parsimonious statistical model from among 11 commonly used models on a gene-by-gene basis to capture each gene's expression distribution.
• Parameter Estimation: M3S estimates model parameters within the chosen model framework to accurately represent underlying gene expression patterns.
• Differential Gene Expression Testing: M3S performs differential gene expression analysis using the selected models to identify genes with significant expression changes across conditions or cell states.

Scientific Applications:

• Single-cell transcriptomics: Modeling multimodal gene expression distributions in scRNA-seq data to support interpretation of heterogeneous cell populations.
• Bulk tissue transcriptomics: Applying model selection and parameter estimation to large-scale bulk tissue transcriptomic data to capture multimodal expression patterns.
• Characterization of cell heterogeneity: Identifying and describing heterogeneous cell types or states within complex tissues by capturing multimodal expression profiles.

Methodology:

Selects among 11 statistical models on a gene-by-gene basis, estimates parameters for the selected model, and conducts differential gene expression testing using the selected models.