Computationally efficient experimental design strategy for reducing gene network uncertainty

Abstract

In this work, we present a computationally efficient method for selecting experiments that can effectively reduce the uncertainty in gene regulatory networks (GRNs). The proposed method prioritizes potential experiments based on the mean objective cost of uncertainty (MOCU) that is expected to remain after performing the experiments. A network reduction scheme is used to approximately estimate the MOCU at a reduced computational cost without disrupting the ranking of potential experiments. The effectiveness of our method is demonstrated through simulations.