A0250
Title: Minimax optimum design for choosing between models for enzyme inhibition
Authors: Ellinor Fackle-Fornius - Department of Statistics (Sweden) [presenting]
Abstract: In enzyme kinetics when studying inhibition, an extended version of the Michaelis-Menten model can be used to model two different types of enzyme inhibition: competitive and non-competitive inhibition. In order to discriminate between the two types of inhibition precise estimation of one of the model parameters is desired and a DS-optimum design would be suitable. Since the optimum design will be parameter dependent one option is to use the minimax principle for design optimality. The minimax design seeks the minimum of the maximum criterion function for a set of plausible parameter values specified by the researcher. Thus, it aims to be robust as long as the parameter values belong to the prespecified set. In this application, we derive minimax DS-optimum as well as maximin efficient designs (where design efficiency in relation to the locally optimal design is the criterion). We evaluate the minimax and maximin efficient designs in terms of variance and efficiency over the prespecified set of parameter values and make comparisons with locally optimum designs and another type of robust design.
