A fuzzy logic approach for regulation in flux balance analysis

Abstract

Technological advances in experimental observations give the ability to examinecomplex biological systems and the chance to forecast the bacterial behaviour in silico. In this thesis, the experimentally observed behavior by which bacteria are able to establish a timehierarchy of sugar utilization is examined and simulated. The optimal growth of bacteria ondefined carbon sources are now easily predicted from the solutions of constraint basedmetabolic models. However, these methods are unable to predict the sequence of carbon utilization and changes in cellular behavior of growth in mixed substrates. In this work aregulatory structure describing transcriptional regulation of catabolic genes or operonsexpressed in Fuzzy Logic Formalism is combined with dynamic Flux Balance Analysis(FBA). Since the transcription of operons is regulated by specific promoters and inducers that evolve from substrate usage, this regulatory structure is a natural part of any model of carbonsource utilization. The Fuzzy Logic Formalism is a good alternative to differential equationmodels that require kinetic parameter values and superior to Boolean Formalism whichautomatically sets regulation as "on" or "off" rules. The FBA/Fuzzy Logic combination was successfully used to simulate aerobic growth ofEscherichia coli in mixed double (glucose-lactose) or triple (glucose-lactose-galactose, glucose-sorbitol-glycerol) substrates and anaerobic growth of Lactococcus lactis in a triplesubstrate (glucose-lactose-galactose). When well-defined data are available, the computed results are in good agreement with the data. The method also allows for the prediction ofgrowth lag periods upon substrate substitution and changes in growth pattern and substrateutilization upon pulse injection of substrates in existing growth media.