DMET-miner: Efficient learning of association rules from genotyping data for personalized medicine

Recent developments of microarray technology enable the investigation of allelic variants that may be correlated to phenotypes. In particular the Affymetrix DMET (Drug Metabolism Enzymes and Transporters) platform enables the simultaneous investigation of all the genes that are related to drug absorption, distribution, metabolism and excretion (ADME) and it has been used in clinical studies. In a previous work we developed DMET-Analyzer, a platform able to automatize the study of allelic variants, that has been validated in clinical studies. DMET-Analyzer is able to correlate a single variant for each probe (related to a portion of a gene) through the use of the Fisher test, on the other hand it is unable to discover multiple associations among allelic variants. To overcome those limitations, here we propose DMET-Miner, that is able to correlate the presence of a set of allelic variants by employing an Apriori-like discovery strategy. Preliminary experiments on a synthetic DMET dataset.