Research over the past 30 years has elucidated the roles of polymorphic human liver cytochrome P450 (P450) enzymes associated with toxicological and/or pharmacological actions. Thalidomide exerts its various pharmacological and toxic actions in primates through multiple mechanisms, including nonspecific modification of many protein networks after bioactivation by autoinduced human P450 enzymes. To overcome species differences between rodents, currently, nonhuman primates and/or mouse models with transplanted human hepatocytes are used. Interindividual variability of P450-dependent drug clearances in cynomolgus monkeys and common marmosets is partly accounted for by polymorphic P450 variants and/or aging, just as it is in humans with increased prevalence of polypharmacy. Genotyping of P450 genes in nonhuman primates would be beneficial before and/or after drug metabolism and toxicity testing and evaluation as well in humans. Genome-wide association studies in humans have been rapidly advanced; however, unique whole-gene deletion of P450 2A6 was subsequently developed to cover nicotine-related lung cancer risk study. Regarding polypharmacy, toxicological research should generally be aimed at identifying the risk of adverse drug events following specific potential drug exposures by examining single or multiple metabolic pathways involving single or multiple drug-metabolizing enzymes. Current and next-generation research of drug metabolism and disposition resulting in drug toxicity would be addressed under advanced knowledge of polymorphic and age-related intra- and/or interspecies differences of drug-metabolizing enzymes. In the near future, humanized animal models combining transplanted hepatocytes and a humanized immune system may be available to study human immune reactions caused by human-type drug metabolites. Such sophisticated models should provide preclinical predictions of human drug metabolism and potential toxicity.