SNP Aassociation Study By Genetic Particle Swarm Optimization

Genome-wide association study is a powerful way to analyze the relationship between complex diseases and genes, single nucleotide polymorphism(SNP) is the main tool for the genome-wide association analysis. Through correlation analysis between SNP and disease, we can quickly identify susceptibility loci for complex diseases. However, due to the huge number of SNPs and the relatively small sample size, coupled with the impact of the epistasis, the space of searching is amazing and making complex disease susceptibility loci difficult to detect above.In This article, I mainly refer to research the association between complex disease and the SNPs by GPSO. After analysis the good point of GA and PSO, I introduce the GA in PSO when the PSO update the speed of the particle swarm.This article use three ways to updata the speed: weak association,strong association and better association. The strong association is refers to in the process of location updating we tend to look for a single SNP and disease relevance of the site. While weak association is refers to in the process of location updating we tend to find a single SNP and disease association weaker sites.In the better association in association analysis we update at all possible locations to find the optimal location of exhaustion. By combining particle swarm algorithm and genetic algorithm approach hand overcomes the defects of slow convergence of genetic algorithm at first, on the other hand is the particle swarm optimization algorithm of excellent global search ability is introduced, greatly enhance the time efficiency of the algorithm. This article first conducted experiments on simulated data sets validate the feasibility and advantages of the algorithm, followed by another test on real data, to identify potential loci. By comparison and analysis the difference experimental results,we can conclude that the algorithm has a better performance during the global search.