DSM and Delft University of Technology (TU Delft) are jointly setting up the Artificial Intelligence Lab for Biosciences (AI4B.io Lab), as part of the Dutch National Innovation Center for AI (ICAI). AI is already widely applied in engineering research but this lab is billed as the first to explore AI’s potential in biosciences and biotechnology.
AI turns the traditional trial and error approach to scientific research on its head. The desired objective is brought to life in a digital environment using ‘digital twins’ (a virtual mirror of the desired real world situation), while machine learning helps determine how to achieve it. In this way, innovation cycle time is reduced.
Biotechnology is used to conceive a range of food ingredients and products, from baking enzymes to gluten-free bread and beer, meat substitutes, lactose-free milk and sugar-reduced beverages. By applying AI, DSM says it is aiming for more efficient operation of its processes, but also improved quality control of its products – for example through particle size modelling and control during product formulation processes like drying.
From microbe to factory scale
The lab will apply AI to biotechnology and biomanufacturing, from microbial strain development to process optimisation and scheduling.
“The AI4B.io Lab is envisioned to speed up innovation, for example through the development and application of highly accurate AI based models of microbes, bioprocesses and the combination of both. An important aspect of the Lab is the exploration of AI methods all the way from microbe-level to a full factory scale, with applications ranging from strain development and application development to fermentation process control and scheduling,” Marcus Remmers, chief technology officer at DSM, told this publication.
Remmers explained how AI can help biomanufacturing innovation and full-scale operation in multiple ways.
“One way is to find patterns in data to guide optimisation processes in R&D, for example, in automated strain characterisation and development projects,” he said.
Another is speeding up calculations, where AI-based models allow for real-time optimisation and adaptive steering of processes based on highly accurate models, so called ‘digital twins’.
“These take into account microbial behaviour in a full-scale bio-reactor while simulating flow patterns and oxygen and sugar gradients and acting on real-time process-sensor data,” explained Remmers.
AI in action
Taking the example of fermentation steering, in most industrial settings today, fermentation relies heavily on the intuition and experience of operators, who decide how long to run the project, at what speed to feed the microorganisms, and how to regulate the temperature and air supply. Based on insights from historic runs, an Artificial Intelligence (AI) based system could predict outcomes of different parameter combinations and propose best parameter steering.
The AI4B.io Lab doesn’t have any projects underway yet, but when it is up and running, digital twin concepts and the use of AI to optimise fermentation processes, factory scheduling or metabolic pathway analysis will be some of the projects it will be working on, according to Remmers. Scientists there will also be focusing on using AI methods to understand the microbiome and advanced DNA analysis.
Led by Professor Marcel Reinders, director of TU Delft Bioengineering Institute, the lab’s workload will combine commercially-driven and public interest projects, confirmed Remmers.
“We use DSM test cases and real world problems to build a deeper understanding of underlying mechanisms. By solving these problems we learn more about potential applications of AI, its strengths and limitations These insights are also benefitting the ICAI institute overall,” he said.
Timing-wise, the lab has just started the process of hiring PhDs and is expecting first results early in 2021.
DSM has committed to investing €2.5 million in the laboratory over the first five years, on the basis that it regards biosciences as an important tool for addressing climate change and resource scarcity and optimising the global food system.
“Biotechnology allows us to use microorganisms – like algae, fungi and bacteria – to produce molecules, such as vitamins or lipids, from raw materials like sugars and plant extracts. Through biomanufacturing we can turn these molecules into sustainable biobased products and solutions, like dairy alternatives of plant-based protein solutions. This way, biotechnology and biomanufacturing help us to sustain and restore healthy populations,” said Remmers.