Genome-Scale Metabolic Model's multi-objective solving algorithm based on the inflexion point of Pareto front...

The solution of Genome-Scale Metabolic Model (GSMM) directly affects the simulation accuracy of the metabolic process in digital cells. Single-objective optimization methods, such as flux balance analysis (FBA), which is widely used in solving GSMM, have limitations when simulating actual biological processes, which leads to unrealistic results due to other biological constraints being ignored. A novel multi-objective differential evolution algorithm based on general FBA (i.e., differential evolution FBA [DEFBA]) is hence proposed to solve GSMM. First, in accordance with the assumption that cells minimize resource consumption and maximize resource utilization, the maximum specific growth rate and the minimum cellular production rate of ATP, NADPH, and NADH are defined as the multi-objective functions of DEFBA. Second, FBA is used to produce the initial individuals of DEFBA by changing the upper bound of biomass reaction in GSMM. Third, mutation and selection operations help in generating new individuals in the solution space to search the Pareto front. Finally, the optimal solution is selected by analyzing the inflexion point of the Pareto front. In DEFBA, multi-objective technology and optimal solution judging technology can introduce the biological constraints into the GSMM solving method, such that the solution can be more consistent with the essential biological mechanism. DEFBA is applied to solve Aspergillus niger's GSMM. The improved results show that DEFBA can be an effective general solving algorithm for GSMM.