D3GRN

D3GRN infers gene regulatory networks from gene-expression data to elucidate regulatory mechanisms underlying cellular processes and disease by constructing dynamic networks and decomposing regulatory relationships into functional sub-problems.

Key Features:

• Dynamic Network Construction: Models temporal changes and interactions within gene regulatory networks (GRNs) through dynamic network construction.
• Data-Driven Functional Decomposition: Transforms GRN inference into a functional decomposition problem by breaking each target gene's regulatory relationships into sub-problems.
• Algorithm for Revealing Network Interactions (ARNI): Applies ARNI to solve sub-problems and identify candidate interactions from gene-expression data.
• Bootstrapping and Area-Based Scoring: Uses bootstrapping combined with an area-based scoring strategy to refine inferred networks and address limitations of constructing networks solely from unit-level data.

Scientific Applications:

• GRN inference: Inferring gene regulatory networks from gene-expression data to map regulatory relationships among genes.
• Benchmarking and performance evaluation: Validation and comparison on DREAM4 and DREAM5 benchmark datasets with evaluation by Area Under the Precision-Recall Curve (AUPR).
• Systems biology and disease research: Elucidating complex regulatory mechanisms across biological contexts to inform studies of cellular processes and disease mechanisms.

Methodology:

Performs dynamic network construction; decomposes regulatory relationships for each target gene into sub-problems (functional decomposition); applies the Algorithm for Revealing Network Interactions (ARNI) to each sub-problem; and refines inferred networks using bootstrapping and an area-based scoring method.