Compartmental modeling framework for investigating pathway-based statistical methodology and study design in candidate gene association studies

Many statistical methods have been introduced for candidate gene association study of complex diseases. However, these methods typically used a simple penetrance function linking genetic variation to phenotype without incorporating detailed biologic information. We developed a physiologically based compartmental model (CM) to simulate glucose and insulin profiles following oral glucose tolerance test, in which each model parameter directly links to a specific biologic function. Then, "genes" can be assigned to model parameters to simulate specific genetic defects in the glucoregulatory system. Through simulations, we demonstrate that CM provides helpful information for study designs and pathway-based statistical analysis. This application represents a unique combination of statistical genetic and compartment modeling methodology, and provides an excellent simulation tool for novel statistical genetic methods to untangle the genetic architecture of complex diseases.