A comparative analysis of chaotic particle swarm optimizations for detecting single nucleotide polymorphism barcodes

Area of Science:

• Computational biology
• Bioinformatics
• Statistical genetics

Background:

• Large-scale genetic datasets pose computational challenges for statistical analysis.
• Single nucleotide polymorphism (SNP) barcodes are crucial for disease association studies.
• Previous chaotic particle swarm optimization (CPSO) methods have shown promise in SNP barcode detection.

Purpose of the Study:

• To evaluate additional chaotic maps integrated with particle swarm optimization (PSO) for enhanced SNP barcode detection.
• To assess the efficacy of various CPSO methods in analyzing high-dimensional genetic data for disease association.

Main Methods:

• Nine distinct chaotic maps were incorporated into the PSO algorithm.
• Comparative analysis of CPSO methods using XOR and ZZ disease models.
• Efficacy evaluation based on chi-square (χ²) test statistics and simple linear regression of gbest values.

Main Results:

• Chaotic maps significantly improve PSO's ability to escape local optima.
• The Sinai chaotic map demonstrated superior performance, yielding the highest χ² values and identifying optimal SNP sets.
• Sinai chaotic map-PSO achieved high β values (≥0.32 for XOR, ≥0.04 for ZZ) and significant p-values (<0.001).

Conclusions:

• The Sinai chaotic map effectively enhances PSO's fitness values (χ²) for SNP barcode detection.
• The proposed Sinai chaotic map-PSO method is highly effective for identifying potential SNP barcodes in disease models.
• This approach offers a computationally efficient solution for analyzing large genetic datasets in disease research.