Cascade Gaussian Process Regression Framework for Biomass Prediction in a Fed-batch Reactor

Biotechnological production of omega-3 fatty acids: current status and future perspectives

“…Different ML techniques have been employed to manage fermentation parameters such as temperature, pH, dissolved oxygen, aeration rate, feeding rate, and agitation speed. For instance, Gaussian process regression (GPR) was implemented to forecast the biomass concentration of S. cerevisiae based solely on substrate flow rate, given that certain variables could not be measured in real-time (Masampally et al, 2018). Similarly, the challenge of optimum growth temperature (OGT) detection in non-conventional microorganisms has been solved using the amino acid composition as input for employing six different ML models since amino acid composition is strongly related to the OGT (Li et al, 2019).…”

Metabolomics and modelling approaches for systems metabolic engineering

“…ANN was further used in conjunction with a genetic algorithm to determine the best nutrient composition for producing potential anti-cancer compounds in S. cerevisiae ( Zheng et al, 2017 ). In a different study, the focus is on predicting the required concentration of S. cerevisiae in a bioreactor, by simply using substrate flow rate as input ( Masampally et al, 2018 ). To be able to do so, three sequential models, using Gaussian process regression based on Bayes’ rules, were applied to infer important intermediate variables, such as the gas hold-up, and the concentrations of biomass and dissolved oxygen, which could not be measured in real time.…”

Machine learning for synthetic biology: Methods and applications

“…代谢工程结合 酶学、化学计量学、分子生物学、数学和组学等研究 手段, 对细胞代谢途径进行修饰和调控, 实现特定目的 产物的高效合成. 在随后的几十年中, 代谢工程领域发 展迅速, 不仅在生物发酵, 同时在医药、植物、动物、 环境和疾病治疗等领域均有重要应用 [114,115] ; 用于代谢途径构建的3N-MCTS [8] ; 用于代谢流 调控的GEMs [116] 、MiYA [117] 及Dcell [118] ; 用于扩大化培 养的SVM [119] 、Fuzzy logic [120] 和GPR [121] . 率、精确度及脱靶效应等方面仍然面临诸多挑战 [125] .…”