CENA

CENA identifies dynamic, subset-specific pairwise associations between gene expression from single-cell RNA sequencing (scRNA-seq) and additional cellular metadata to characterize context-dependent gene relationships.

Key Features:

• Subset-Specific Association Detection: Detects pairwise gene expression associations that are confined to particular subpopulations of cells rather than uniformly across the entire population.
• No Predefined Subsets Required: Predicts cell subsets based on shared characteristic states instead of requiring predefined cell annotations.
• Dynamic Modulation Analysis: Reveals how gene dependencies change over time and across cellular states to capture temporal and state-dependent modulation.
• Scalability: Demonstrated scalability to large scRNA-seq datasets.
• Validation through Simulated Data: Performance validated on simulated datasets, where it outperformed existing methods in identifying dynamic associations while maintaining scalability.
• Application to Real Biological Data: Recovers dynamic changes in gene associations in real scRNA-seq datasets that can be masked by averaging across heterogeneous cell populations.

Scientific Applications:

• Developmental Biology: Dissects context-specific and temporal gene interactions during development by identifying subset-specific associations.
• Cancer Progression: Identifies dynamic, context-dependent gene associations relevant to tumor heterogeneity and progression.
• Response to Treatment: Detects treatment-associated changes in gene relationships across cellular subpopulations.

Methodology:

Analyzes scRNA-seq data alongside additional cellular metadata to detect subset-specific pairwise associations, predicts subsets based on shared cellular states, and identifies dynamic dependencies along cellular trajectories; validated using simulated datasets.