Automatic generation of a digital twin for continuous antibody capture

Digitalization of the downstream process for monoclonal antibodies can be used as a tool towards more efficient and sustainable processing. One part of digitalization are digital twins, which can help provide insights into a process and be used for process optimization, reducing production costs and accelerating drug discovery. However, creating a useful digital twin is often difficult and time-consuming. To facilitate the acquisition of useful digital twins, we developed a procedure for automatically generating and calibrating a digital twin for a continuous 3-column periodic countercurrent chromatography capture step of monoclonal antibodies. The procedure was conducted in two segments. First, the digital twin was automatically generated, based on the configuration of the setup. Second, a predefined calibration procedure, including calibration experiments and simulations, was automatically executed, calibrating the digital twin. To test its value, the digital twin was used for two separate optimizations: one to optimize yield and resin utilization and another to account for differences in column capacities when using a column with reduced capacity. The result from the first optimization was validated in an experiment, resulting in a 0.6 % difference in yield and resin utilization compared with the simulation. The second optimization increased resin utilization by 1.2 %, versus the first optimization, at a yield of 99.8 %. Our procedure can be used to generate and calibrate digital twins efficiently, constituting a step toward integrating digitalization into the downstream process for monoclonal antibodies, which could lower production costs and increase sustainability.