Seeding-inspired chemotaxis genetic algorithm for the inference of biological systems

Area of Science:

• Systems Biology
• Computational Biology
• Bioinformatics

Background:

• Post-genomic research requires quantitative, dynamic interaction data for biological systems.
• S-systems are promising models for gene/protein interactions but face challenges in parameter and structure identification.
• Existing evolutionary computation methods struggle with high-dimensional systems and efficient convergence.

Purpose of the Study:

• To develop an advanced optimization algorithm for identifying biological system structures and parameters.
• To address limitations in current methods for high-dimensional gene regulatory network analysis.
• To improve the speed and accuracy of network inference in systems biology.

Main Methods:

• Introduction of a seeding-inspired chemotaxis genetic algorithm (SCGA).
• SCGA utilizes seeding-inspired genetic operations for enhanced exploitation and exploration.
• Employs winner-chemotaxis-induced population migration for adaptive search and convergence.

Main Results:

• SCGA successfully identified correct biological network structures within 1-3 pruning steps.
• Demonstrated high pruning safety and accuracy, with truncated term values below 10^-14.
• Effective performance on thirty-gene systems, showcasing scalability for high-dimensional problems.

Conclusions:

• SCGA offers a robust and efficient solution for inferring complex biological network structures.
• The algorithm overcomes limitations of existing methods in handling high-dimensional systems.
• SCGA facilitates accurate and safe identification of gene/protein interactions in systems biology.