Discrimination of Metabolic Flux Profiles Using a Hybrid Evolutionary Algorithm

Abstract

Studying metabolic fluxes is a crucial aspect of understanding biological phenotypes. However, it is often not possible to measure these fluxes directly. As an alternative, fluxome profiling provides indirect information about fluxes in a high-throughput setting. In this paper, we consider a scenario where fluxome profiling is used to investigate characteristic differences between a number of bacterial mutant strains. The goal is to identify groups of mutants that show maximally different fluxome profiles. We propose an evolutionary algorithm for this optimization problem and demonstrate that it outperforms alternative methods based on principle component analysis and independent component analysis on both real and synthetic data sets.