The Intensified Design of Experiment (iDoE) approach has the potential to significantly reduce the time needed for process development and characterization.
“We’ve proved iDoE on a 20 L scale and compared it to a full design space involving three parameters set at three different levels,” explains Mark Dürkop, PhD, from the University of Natural Resources and Life Sciences in Vienna and CEO of Novasign.
One bottleneck in improving the quality of innovative biopharmaceuticals is understanding cell culture growth and product formation in bioreactors. According to Dürkop, this characterization process takes a long time because there are numerous parameters, such as pH, feeding rate, aeration rate, or temperature, affecting the quantity and quality of the product.
“If there are three critical process parameters each on three levels, 27 experiments have to be performed to understand and screen the full design space,” he explains. “If one cultivation takes two weeks and you only have a single bioreactor, that’s an entire year spent doing experiments.”
An alternative is an iDoE approach whereby, instead of keeping parameter combinations constant for every experiment, multiple parameters are changed within a single cultivation, Dürkop explains. A machine-learning algorithm is then used to determine the effects of each parameter. The model is further applied in a digital twin application to understand cell behavior in more detail.
The team decided to test the iDoE approach for process characterization of E. coli in a 20 L bioreactor. The E. coli had been engineered to express a recombinant enzyme being researched as a treatment for neonatal lung injury.
The intensified approach described the process design space in just nine experiments, 66% faster than a classic Design of Experiment (DoE) approach.
“This approach significantly speeds up characterization times,” explains Dürkop. “You could also use it during process development, but characterization is ideal because you don’t deviate too much from operational conditions.”
A paper covering this research was recently published in Biotechnology Journal online.
The post AI Promises to Accelerate Process Characterization appeared first on GEN – Genetic Engineering and Biotechnology News.