DEMA

DEMA integrates quantitative biological properties into graph layout algorithms to produce biomolecular network visualizations that reflect interaction coefficients, effect coefficients, fold changes in gene expression, and protein–protein interaction weights.

Key Features:

• Distance-bounded Energy-field Minimization Algorithm: Positions nodes using a distance-bounded energy-field minimization approach to encode quantitative constraints in the layout.
• Parameterized Energy Model: Uses a parameterized energy model where nodes are influenced by repulsive and attractive forces derived from network topology and biological parameters.
• Interaction Coefficient: Encodes the strength of gene-to-gene associations and modulates force terms in the layout model.
• Effect Coefficient: Represents relative contribution weights of genes that alter node influence within the energy model.
• Fold Change in Gene Expression: Incorporates fold change values from gene expression data as gene weight inputs to the layout.
• Gene Weights and Functional Properties: Aggregates gene weights alongside protein–protein interaction weights, gene-to-gene correlations, and gene set annotations as four parameterized functional properties.
• Attraction/Repulsion/Grouping Coefficients: Applies attraction, repulsion, and grouping coefficients to generate customizable network views based on biological parameters.

Scientific Applications:

• Gene candidate discovery: Supports identification and prioritization of gene candidates within complex biomolecular networks by embedding quantitative measures into spatial layouts.
• Disease network analysis: Applied to genetic data from autism spectrum disorder and Alzheimer's disease to demonstrate interpretability and candidate prioritization in disease-relevant networks.

Methodology:

Implements a parameterized energy model in which nodes experience repulsive and attractive forces derived from network topology and biological parameters (interaction coefficient, effect coefficient, fold change), with gene weights, protein–protein interaction weights, gene-to-gene correlations, and gene set annotations forming four parameterized functional properties and explicit attraction/repulsion/grouping coefficients.