Skip to Main Content
The progress in detecting single nucleotide polymorphisms (SNPs) genome-wide provides the opportunities to investigate responsible loci for multigenic human disorders, simultaneously proposes the urgent need to establish the technology for large-scale analysis of SNPs. In this study, a heuristic approach is proposed to mine SNP markers responsible for disease. Firstly, a feature selection algorithm is advanced accounting for both joint and marginal effects. Secondly, we iterate the algorithm to identify plentiful subsets of SNP markers which have potential to discriminate between affected sib-pairs with disease and unaffected controls based on the proportion of alleles identical by descent (IBD) at the SNP locus, for sibling pairs. Those markers that are returned most often from many subsets sampled are considered "important". We have applied the approach to the Genetic Analysis Workshop 14 COGA data and some intriguing results emerge.