| Metabolic pathways are candidates for examining gene-exposure interaction effects on complex diseases. We present a hierarchical modeling platform integrating physiologically based pharmocokinetic models and Bayesian approaches to characterize gene-exposure interaction effects in a metabolic pathway. We simulate datasets under a known metabolic model. Individual-specific metabolic rates are described by Michaelis-Menten parameters V max and Km, regressed on genotypes at the relevant loci. Intermediate measurements such as enzyme activities and metabolite concentrations are incorporated as flawed measurements of an individual's long-term metabolic rates and metabolite concentrations. Parameter posteriors are estimated using Markov Chain Monte Carlo methods. The platform allows investigation posterior estimation precision relative to the simulated effect magnitude and appears to have great utility for characterizing etiologic genetic and exposure effects and dissecting complex pathways.; Keywords. Metabolic pathways, Physiologically based pharmacokinetic models, Bayesian approaches, Hierarchical modeling, Exposure, Genotype, Michaelis-Menten, Intermediate measurements, Enzyme activities, Metabolite concentrations, Markov Chain Monte Carlo. |
