A Genetic Programming Algorithm for Association Studies
Robin Nunkesser · 2008
Abstract. The analysis of genetic association is useful for identifying genetic fac-tors that may contribute to a medical condition. An important subarea are case-control studies on single nucleotide polymorphism (SNP) data, i.e. data on genetic variations that occur when different base alternatives exist at a single base pair position. The major goal of these studies is to identify SNPs and SNP interactions that lead to a higher disease risk. We present a Genetic Programming algorithm called GPAS (Genetic Program-ming for Association Studies) for case-control association studies which outperforms other regression and discrimination approaches on real and simulated SNP data ex-amples. On these examples, GPAS is also able to identify high-order interactions of SNPs with a high odds ratio, which are not found by other feature selection methods. The algorithm is implemented in an extendible R package called RFreak, which also allows an easy modular implementation of further Genetic Programming and Evolutionary Computation algorithms. Finally, we discuss more areas the algorithm is applicable to and extensions