Predictive Digital Twins Control Bioprocesses for MAb Production

Published on: 

A collaboration between Insilico Biotechnology and IFAT aims to develop a manufacturing, planning, and control system for the production of monoclonal antibodies.

Insilico Biotechnology, a German provider of solutions and software for the simulation of living cells, and Germany’s Institute for Automation Engineering (IFAT) of the Otto von Guericke University Magdeburg are collaborating to develop a manufacturing, planning, and control (MPC) system for the production of monoclonal antibodies (mAbs) in Chinese hamster ovary (CHO) cells. They will combine their modelling, process control, and automation expertise to create the system. The partners announced their collaboration on Aug. 13,2019.

The production of high-quality biologics requires the development of robust and well-understood production processes using mammalian cell cultures. The control of these processes in order to achieve robust product quality and productivity can be significantly improved by online process monitoring followed by corrective actions. Digital twins are virtual representations of the production process which enable preemptive process control by using online data to predict the process outcome in advance. This enables unprecedented possibilities for timely and automated intervention to steer the process results at an early stage.

Insilico digital twins are virtual representations of the actual bioprocess that include a genome-based metabolic network of the cell, a mechanistic model of the process, and an artificial neural network. Fusing these three model components enables simulations of a virtually unlimited number of process scenarios and the advance prediction of outcomes due to process parameter changes, according to the company.

Model predictive control will be based on the Insilico digital twin and online process monitoring to establish open-loop decision support (OPL–DS) for process control. For this purpose, the project partners will develop a softsensor, conceive a robust control strategy, and implement a self-learning system for online process control to be combined with the digital twin.

"The jointly developed solution will for the first time enable true online-control of critical quality attributes such as the glycosylation profile. We are positive that partnering with Rolf Findeisen's leading research group is the key to achieving this goal," said Klaus Mauch, CEO of Insilico Biotechnology, in a company press release.


"Predictive process monitoring and control provides highly valuable decision support to the process operator and enables preemptive process steering to ensure product and process specifications are met. Insilico's state-of-the-art [d]igital [t]wins for biomanufacturing processes play a pivotal role in developing this innovative solution," added Rolf Findeisen, professor at University Magdeburg, in the press release.

“Integrating [d]igital [t]wins for prediction, optimization-based decision support and control, with machine learning approaches allows [us] to handle process uncertainties and variability unavoidable in biotechnological production,” said Dr. Lisa Carius, junior research group leader in the field of smart automation of biotechnological processes at the Laboratory for Systems Theory and Automatic Control, IFAT, in the press release.

Source: Insilico Biotechnology