Résumé
Pharmacogenetics has progressively become a major concern in personalized medicine. The development of modern technologies in genetic testing and cost-effectiveness have rendered genotyping strategies easy to perform comparing to time-consuming phenotyping methods. In oncology, canonical markers such TPMT, DPYD or UGT1A1 are routinely included in clinical practice but their use is still controversial partly because of insufficient genotype to phenotype correlation. The next challenge is to accurately translate genotype-phenotype correlations into clinically useful diagnostics, and clinically useful leads concerning new therapeutic targets. Besides, recent studies have focused on emerging genetic variants of ADME genes such cytidine deaminase (CDA) or pregnane X receptor (PXR) that could be of interest for predicting anticancer drug response or toxicity. The candidate gene approach "metabolism guided" now evolves towards more global strategies thanks to genome resequencing projects such HapMap that have considerably increased our knowledge of genetics variations in humans. Multiplexed genotyping methods make possible the set-up of panels of candidate or tag SNPs for subsequent haplotypic analysis. Last, genome-wide association studies (GWAS) are feasible when large cohort of patients is available to identify new loci associated with drug response or adverse drug reactions or to definitively confirm the role of candidate genetic variations.