TGCnA

TGCnA models temporal gene coexpression networks from transcriptomic time-course data using a low-rank plus sparse decomposition to capture shared and time-specific covariance structure.

Key Features:

• Low-Rank Plus Sparse Framework: Jointly models multiple covariance matrices across time points by decomposing each covariance into a low-rank component that captures network similarity across time and a sparse component that captures time-specific changes.
• Scalability: Scales to large gene networks and high-throughput transcriptomic datasets to accommodate many genes and multiple time points.
• Covariance Matrix Estimation: Estimates covariance matrices from time-course gene expression data to infer coexpression patterns beyond conventional correlation-based methods.
• Gene Module Discovery: Identifies gene modules as clusters of genes with coordinated expression over time, including conserved modules and transient interactions.

Scientific Applications:

• Developmental biology: Characterizes dynamic gene network changes across developmental time courses.
• Disease progression studies: Tracks evolution of gene coexpression networks during disease onset and progression.
• Response to treatments or environmental changes: Identifies temporal network responses to interventions or environmental perturbations.
• Biomarker discovery: Detects temporal expression patterns and modules that may serve as biomarkers for phenotypic outcomes.

Methodology:

Input transcriptomic time-course gene expression data; model covariance matrices across time points using a low-rank plus sparse decomposition; analyze resulting networks to identify conserved modules and transient interactions.