GRNsight

GRNsight visualizes gene regulatory networks by displaying genes, transcription factors, and weighted directed regulatory interactions to support analysis of mRNA and protein expression dynamics and weight parameters in GRN models represented by differential equations.

Key Features:

• Input Formats: Parses adjacency matrices provided in Excel (regulators in columns and target genes in rows), Simple Interaction Format (SIF) text files, and GraphML XML files.
• Weighted and Unweighted Networks: Renders both unweighted and weighted directed graphs with visual encodings for activation (positive weights) and repression (negative weights).
• Weight Encoding: Encodes sign via arrowhead shape and color and encodes magnitude via edge thickness.
• Graph Layout: Arranges network topology using a customized force-directed graph layout implemented with D3.js.
• Self-regulation Representation: Represents self-regulatory connections as loops.
• Directed-edge Support: Supports networks composed of directed edges suitable for regulatory interactions.
• Scalability: Optimized for networks with fewer than 35 nodes and 70 edges and capable of handling networks up to approximately 75 nodes or 150 edges.
• Model Focus: Targets visualization of weight parameters from GRN models, including those represented by differential equations in small-scale networks.

Scientific Applications:

• Systems Biology Research: Visualizes regulatory interactions among genes and transcription factors to aid analysis of gene regulatory network structure and dynamics.
• Parameter Estimation: Supports inspection of weight parameters used in GRN models for parameter estimation workflows.
• Forward Simulation Studies: Provides visual context for weight parameters in forward simulation studies of small-scale GRN models.
• Education: Facilitates exploration and explanation of GRN concepts and regulatory motifs such as self-regulation.
• Small Directed Networks in Other Domains: Applies to small directed networks beyond systems biology where directed weighted interactions are analyzed.

Methodology:

Parses adjacency matrices, SIF, and GraphML inputs; implements a customized force-directed layout using D3.js; encodes weight sign by arrowhead shape and color and weight magnitude by edge thickness; visualizes self-regulatory loops and weight parameters for GRN models represented by differential equations.