Application of approximate pattern matching in two dimensional spaces to grid layout for biochemical network maps

Area of Science:

• Computational Biology
• Bioinformatics
• Network Visualization

Background:

• Visualizing large-scale biochemical networks requires efficient node coordinate calculation for clarity.
• Grid layouts offer compact and orderly network maps but often involve high computational costs.
• Existing algorithms struggle with complex constraints, leading to slow performance.

Purpose of the Study:

• To develop a computationally efficient grid layout algorithm for large-scale biochemical networks.
• To improve the speed and effectiveness of network map visualization.
• To preserve topological relationships during layout optimization.

Main Methods:

• A hybrid approach combining a fast, non-grid layout preprocessor with an approximate pattern matching algorithm.
• Distribution of preprocessed nodes onto a square grid using pattern matching.
• Evaluation of the algorithm based on calculation time, crossings, edge lengths, and F-measures.

Main Results:

• The proposed hybrid algorithm demonstrates significantly improved performance compared to existing grid layout methods.
• Achieved outstanding results in terms of calculation speed and layout quality.
• Effectively redistributes nodes on a grid while maintaining network topology.

Conclusions:

• The hybrid grid layout algorithm offers a breakthrough in network visualization by efficiently distributing nodes.
• Approximate pattern matching enables rapid layout adjustments while preserving network structure.
• The developed application program is available for free download.