Item

Abstract

Author summary Enzymes play a crucial role in metabolic processes, and by selecting enzymes with optimal activity, we can enhance the production of desired compounds. However, there has been no computational approach to designing metabolic engineering strategies based on the modification of enzyme activities. In this study, we developed a new computational method called Overcoming Kinetic rate Obstacles (OKO) to increase chemical production by modifying enzyme activities in E. coli and S. cerevisiae. Our method uses growing data on enzyme efficiency from experiments and deep learning models to suggest strategies for metabolic engineering. We applied OKO to increase the production of over 40 different compounds in models of E. coli and S. cerevisiae, and found that it can at least double their production without severely affecting cell growth. Finally, we showed that by refining OKO to account for changes in both enzyme activity and abundance, our method can more effectively use existing data and models to design precise strategies for improving chemical production. Our findings pave the way for advanced metabolic engineering techniques for various biotechnological applications.