Nonlinear state estimation as tool for online monitoring and adaptive feed in high throughput cultivations

Abstract

Robotic facilities that can perform advanced cultivations (e.g., fed‐batch or continuous) in high throughput have drastically increased the speed and reliability of the bioprocess development pipeline. Still, developing reliable analytical technologies, that can cope with the throughput of the cultivation system, has proven to be very challenging. On the one hand, the analytical accuracy suffers from the low sampling volumes, and on the other hand, the number of samples that must be treated rapidly is very large. These issues have been a major limitation for the implementation of feedback control methods in miniaturized bioreactor systems, where observations of the process states are typically obtained after the experiment has finished. In this work, we implement a Sigma‐Point Kalman Filter in a high throughput platform with 24 parallel experiments at the mL‐scale to demonstrate its viability and added value in high throughput experiments. The filter exploits the information generated by the ammonia‐based pH control to enable the continuous estimation of the biomass concentration, a critical state to monitor the specific rates of production and consumption in the process. The objective in the selected case study is to ensure that the selected specific substrate consumption rate is tightly controlled throughout the complete Escherichia coli cultivations for recombinant production of an antibody fragment.

Enzyme-Mediated Exponential Glucose Release: A Model-Based Strategy for Continuous Defined Fed-Batch in Small-Scale Cultivations

“…The cells were separated from the supernatant by centrifugation at 15,000× g for 10 min, and the concentration of glucose, acetate and magnesium were determined in the supernatant using the Cedex Bio HT Analyzer. An additional estimate of the biomass concentration X NH 3 was obtained from the added ammonia [33]. The amount of the recombinant product was obtained from fluorescence measurements based on a previously determined conversion factor.…”

“…The model-based adaptive input design was implemented using an in-house developed Python-based framework for simulation [33] with symbolic differentiation [37], and parameter estimation using lmfit [38] and pygmo [39]. For lmfit the method "least_squares", and for pygmo the algorithm "differential evolution" ("de1220") with a population size of 20 and generation number of 200 was used.…”

“…Please refer to the section "Calculation of dilution by bolus additions" in the Supplementary Materials for equations and explanation. Further details on the simulation and parameter fitting procedure are given in Kemmer et al [33].…”

“…The results from the kinetic experiments were validated in parallel high-throughput MBR cultivations with enzymatic glucose feed. A previously described microbial growth model [33] was extended with the mechanistic model for the enzyme-mediated glucose release and was fitted to the validation experiments with enzymatic glucose feed.…”

“…After the evaluation of different mechanistic starch hydrolysis models presented in literature, the model by Polakovič and Bryjak [32] was selected and adapted under conditions relevant for microbial growth. The model was embedded into a modeling and bioprocess control framework [33], which allows the prediction of the glucose release and the progression of the microbial cultivation. The additions of enzyme needed to achieve specific glucose release rates were calculated and automatically realized during parallel MBR cultivations by a liquid handler.…”