Graphics Processing Unit-Enhanced Genetic Algorithms for Solving the Temporal Dynamics of Gene Regulatory Networks

Area of Science:

• Computational Biology
• Bioinformatics
• Systems Biology

Background:

• Gene expression regulation is crucial in biology.
• Determining temporal dynamics of gene networks is challenging due to computational complexity.
• Genetic algorithms offer a promising approach for solving these complex problems.

Purpose of the Study:

• To develop and evaluate efficient parallel implementations of genetic algorithms on Graphics Processing Units (GPUs) for analyzing gene regulatory networks (GRNs).
• To investigate various parallel genetic algorithm schemes and selection methods for optimizing computational performance and solution accuracy.
• To provide guidance for leveraging GPU parallelization in computational biology and bioinformatics.

Main Methods:

• Implementation of parallel genetic algorithms on GPUs using CUDA.
• Systematic study of master-slave, island, cellular, and hybrid models.
• Evaluation of selection methods including roulette and elitist selection.
• Optimization of GPU resource utilization for enhanced performance.

Main Results:

• The island model with elitist selection demonstrated superior performance and genetic algorithm fitness.
• This optimized approach successfully identified the optimal solution for the analyzed gene regulatory network.
• A multifold speedup was achieved on a medium-class GPU compared to a sequential CPU implementation.

Conclusions:

• Parallel genetic algorithms on GPUs, particularly the island model with elitist selection, significantly accelerate the analysis of gene regulatory network dynamics.
• This approach balances finding optimal solutions with maintaining genetic diversity, leading to effective problem-solving.
• The findings offer practical insights for researchers applying nature-inspired metaheuristic algorithms to complex biological problems on massively parallel hardware.