Knowledge of cell physiology is important for making process control decisions in the production of therapeutic proteins. In the current cell culture paradigm, analytical techniques used to obtain data requires substantial effort associated with sampling and off-line experimentation and analysis. An ideal alternative approach is decision making based on robust in-silico prediction of cell physiology that virtually eliminates experimental effort and offers a seamless and real-time approach for process control.
Agent-Based Modeling (ABM), is a novel modeling approach to address the complexity of systems that comprise heterogeneous interacting individuals. An agent is an autonomous computational entity that takes actions within its environment based on the conditions it perceives. We have integrated agent-based modeling of cells in a bioreactor with high-performance computing to leverage parallel processing that allows the ABM program to run faster, more efficiently, and with a higher capacity for the amount of cells that can be modeled. Results of this model will be discussed in this presentation.
Note: This work is a collaboration between Amgen, Inc. and Illinois Institute of Technology and one of the coauthor's affiliation with Intech does not give Intech any rights to this technology.
