A Genetic Algorithm to Filter SNPs for SNP Association Study

Recent advances in human and medical genetics have made it widely accepted that complex diseases are speculated to be caused by multiple genetic variants, their interactive effects, and/or their interactions with environment factors. To find causative factors from genome-wide single-nucleotide polymorphisms (SNPs), SNP filtering is necessary and the key process for follow-up scientific-oriented causative factor discovery and causative interaction modeling. This paper proposes a novel genetic algorithm for SNP filtering. By introducing SNP frequency and using genetic algorithm, experiments on famous trunk data for feature selection, on SNP data for SNP filtering and on real AMD SNP dataset, indicate that the proposed algorithm is potential for both feature selection and feature filtering for genome-wide scale datasets.