CAS Develops New Algorithm to Boost Metabolic Engineering Efficiency

Researchers from the Tianjin Institute of Industrial Biotechnology at the Chinese Academy of Sciences have developed a novel metabolic engineering target design algorithm called ET-OptME. This advancement is significant as metabolic engineering, a fundamental technique in synthetic biology, is increasingly vital for enhancing microbial production of essential compounds. The predictive identification of metabolic targets is a critical first step in the design-build-test-learn cycle, influencing both experimental efficiency and resource allocation. The complexity of cellular metabolism, driven by intricate enzyme-catalyzed reactions and thermodynamic principles, necessitates sophisticated models for accurate predictions. While AI models offer powerful tools, cell mechanism models based on the physical chemistry of metabolic networks provide unparalleled explainability and are crucial for improving the precision of metabolic engineering designs. Traditional algorithms like OptForce and FSEOF rely heavily on stoichiometric models, which fail to account for key physiological mechanisms such as enzyme resource consumption and thermodynamic feasibility. Inspired by nature's "energy-saving and efficient" approach to free energy and enzyme resource coordination, the researchers introduced the ET-OptME framework. This framework integrates enzyme constraints and thermodynamic constraints into the metabolic target design process, enhancing the physiological relevance and practical feasibility of the predictions. The ET-OptME framework consists of two core algorithms: ET-EComp and ET-ESEOF. ET-EComp identifies enzymes that should be upregulated or downregulated by comparing enzyme concentration ranges across different cellular states. ET-ESEOF, on the other hand, scans changes in enzyme concentrations during the process of increasing target flux to capture regulatory signals. Additionally, the framework adopts a "protein-centric" strategy, moving away from traditional reaction-level target prediction and addressing the challenge of controlling multifunctional enzymes. In five industrial case studies involving Corynebacterium glutamicum, ET-OptME demonstrated a 292% improvement in minimum precision over stoichiometric algorithms and a 106% increase in overall accuracy. Compared to state-of-the-art enzyme constraint algorithms, ET-OptME maintained a 70% precision advantage and a 47% accuracy advantage. The study further analyzed the success factors behind key target predictions, such as pyc, gapA, and leuA, highlighting the effectiveness of the enzyme-thermodynamic constraint approach in optimizing pathway efficiency and overcoming metabolic bottlenecks. These findings underscore the potential of ET-OptME in advancing the field of metabolic engineering and improving the efficiency of microbial production processes. The research has been published in the journal Metabolic Engineering and was supported by the National Key Research and Development Program and the National Natural Science Foundation of China. For more details, the full paper can be accessed through the provided link, which includes a schematic diagram of the ET-OptME algorithm.