Artificial intelligence system for enhanced automated 1,3-propanediol green biosynthesis

Abstract

The biorefinery concept holds the promise of green industrial production of biofuel or chemicals, in which fossil resources are substituted by renewable biomass, securing sustainable socioeconomic development. In addition to an optimal microbial cell factory, the fermentation mode is a key factor of biorefinery technology for 1,3-propanediol production from crude glycerol. Although fed-batch fermentation is generally an advantageous mode of submerged fermentation, it requires more sophisticated equipment for online measurement, control techniques for process management, and intelligent decisions during the entire operational process, which are great challenges for the robust and green industrial production of 1,3-propanediol. Here, we developed an extraordinary artificial intelligence system for the entirely automated fed-batch fermentation of 1,3-propanediol, including a sensor, predictor, controller, and automation system. Compared with the constant-speed fed-batch fermentation strategy, the artificial intelligence system automatically regulates the feeding rate and maintains a low concentration of glycerol (∼5 g L⁻¹) and also increases the concentration of 1,3-PDO (64.39 g L⁻¹) and yield (0.58 g g⁻¹) up to 75.7% and 38.1%, respectively. Combined with dynamic metabolic flux analysis, we demonstrate that a low concentration of glycerol controlled by an artificial intelligence system contributes to the balance of the redox pool. An artificial intelligence system for automatic, robust, and enhanced 1,3-propanediol concentration and yield has been successfully developed. This development increases glycerol utilization efficiency and decreases the cost of the medium. It also eliminates the dependence on expensive online instruments and staffing, which may be beneficial for the sustainable biosynthesis of 1,3-propanediol and adaptation to similar production processes.