Optimal Experimental Design (OED) makes use of predictive models to design experiments which increase the knowledge gain while decreasing the needed resources. These methods accelerate bioprocess development, especially in combination with high throughput screening, and thus greatly benefit screening of the best alternative out of a vast variety of strains and cultivation conditions for production optimization. However, the applicability of these tools is low.

Reasons are (1) the lack of suitable and well fitted models, (2) laborious at-line measurements are often needed to generate sufficient quantity and quality data for modelling, (3) missing integration of the product formation into the model and (4) risk of overrunning of cultivation constraints during automatically planned and executed experiments.

AMADEUS tackles these challenges by:

• Reduction of at-line measurements: Soft-sensors/observers will be used to estimate/predict at-line data from available on-line data.
• Implementation of the product formation: (Semi-)automatic product analytics will be developed using the robotic platform in combination with immunoassays and semi-automatic high throughput protein analytics.
• Integration of the product formation in the mechanistic model to include the influence of the cultivation conditions on product quantity and quality.
• Robust OED: Unknown strain properties might lead to overruns of cultivation constraints during robot experiments. To increase the robustness of the process the data will be evaluated, and accumulations of substrate and by-product will be predicted or detected. In this way the violations of cultivation borders will be forecasted and lead to the re-design of the experiment.

The improvement and the applicability of model-based tools in research and industry is the outstanding goal of AMADEUS. The effectiveness of the developed methods will be validated in a case study on the production of antibody fragments in Escherichia coli in an industrial application.

Project partner:
Boehringer Ingelheim (Vienna)

Project duration:
10/2019 - current

Project funding:
Contract researche

Responsible person:
Sebastian Hans
Annina Kemmer