News & Press

Design and Optimization of Enzyme Formulations by Artificial Intelligence

products-right products-right

15 April 2024

Discover the future of enzyme formulation optimization with EnXylaScope’s partner, divis intelligent solutions GmbH, collaborating with Celignis. 

Within the EnXylaScope project, divis intelligent solutions GmbH actively contributes by developing an enzyme formulation optimization tool. This collaborative effort aims to support the design and optimization of enzyme formulations. Project partners, like Celignis, are now using the prototype of this tool to facilitate formulation design.

We pride ourselves on using cutting-edge AI technology to design and facilitate the setup of laboratory experiments for enzyme formulations. Our methods are meticulously tailored to our specific application domain, considering constraints and the desired number of experiments.

After completing laboratory experiments, the system gets performance data from the formulations. Then, it generates a predictive model, enabling accurate forecasts of formulation performance for new enzyme blends. This model allows for precise estimates of how new enzyme blends will best perform.

In EnXylaScope, our application of this approach has yielded remarkable success. The resultant models exhibit high predictive accuracy, enabling us to accelerate the development process in the laboratory substantially.

Ein Bild, das Text, Screenshot, Diagramm, Zahl enthält.

Automatisch generierte Beschreibung

Figure 1: User interface showing a design of experiments (bottom left) for two components (sodium hydroxide, hydrochloric acid) and a first model that has been learned on the experimental data (top right). The two charts show histograms for nitric acid (top left) and calcium hydroxide (bottom right).

Thomas Bäck - divis intelligent solutions GmbH