Improving the Prediction of Phosphate Dynamics in Biotechnological Processes: A Case Study Based on Antibiotic Production Using Streptomyces coelicolor

The objective of this study is to demonstrate that the accurate mathematical description of phosphate dynamics requires a considerable, but unavoidable, degree of complexity when modelling biotechnological systems. As an example, a model predicting antibiotic production using Streptomyces coelicolor is chosen which had difficulties explaining the phosphate dynamics. The model is enhanced by the implementation of an advanced speciation model and a multiple mineral precipitation framework. Furthermore, a model describing intracellular polyphosphate accumulation and consumption is developed and implemented. Based on the conducted work the improved process model is capable of predicting the phosphate dynamics (RMSE≤ 52h: -90 %, RAD≤ 52h: -96 %) very accurately in comparison to the original implementation, where biomass growth was the only phosphate sink. The description of most other variables was improved by a knowledge-based re-estimation of selected parameters as well. This work contributes to the existing process knowledge of biotechnological systems in general and especially to the antibiotic production with S. coelicolor, which emphasizes the necessity of combining physico-chemical and biological processes to accurately describe phosphate dynamics.