Parallel genetic algorithm for disease-gene association

After revealing the complete human genome, computational biology shifted towards the study of associations between complex diseases and genetic markers. Given that the number of human DNA variations, called single nucleotide polymorphisms (SNPs), account for a small percentage of the whole genome, accurate and informative results have become possible. However, a major limitation in association studies is the cost of genotyping SNPs. Therefore, finding a small subset of tag SNPs to be used as good representatives of the rest of the SNPs is essential. In this work, we combine few successful strategies from the literature and present a parallel genetic algorithm for the Tag SNP Selection problem. Our results compared favorably with those of a recognized tag SNP selection algorithm using three different data sets from the HapMap project.