D3E

D3E detects differential gene expression in single-cell RNA sequencing (scRNA-seq) data using a discrete, distributional stochastic model to identify changes in expression patterns and infer average burst size and burst frequency.

Key Features:

• Stochastic model basis: Implements an analytically tractable stochastic model that captures intrinsic variability and noise in scRNA-seq count data.
• Discrete distributional method: Employs a discrete, distributional approach tailored for single-cell RNA-seq data.
• Inference of burst kinetics: Estimates biologically meaningful parameters such as average burst size and burst frequency from expression distributions.
• Detection beyond mean shifts: Detects changes in gene expression patterns even when mean expression remains constant.
• Hypothesis testing: Enables hypothesis testing regarding mechanisms driving changes in gene expression using its stochastic framework.
• Performance evaluation: Was evaluated on synthetic datasets and reported superior performance compared with other methods for identifying differentially expressed genes.

Scientific Applications:

• Single-cell heterogeneity analysis: Characterizes gene expression heterogeneity across individual cells within a population.
• Developmental biology: Supports studies in developmental biology by analyzing single-cell expression heterogeneity.
• Cancer genomics: Supports analysis of tumor heterogeneity and differential expression in cancer genomics.
• Immunology: Supports investigation of single-cell expression differences in immunology.
• Personalized medicine and therapeutic targeting: Contributes to personalized medicine and targeted therapeutic strategies by providing detailed single-cell expression insights.

Methodology:

Uses a discrete, distributional analytically tractable stochastic model to estimate burst kinetics and perform hypothesis testing on expression changes, and was evaluated using synthetic datasets.