Within the discovery-to-market development cycle for biopharmaceuticals, Cell Line Development holds a critical role in creating and selecting a cell line that becomes the Master Cell Bank for clinical and commercial biomanufacturing.  Cell line generation and selection have largely been driven by resource-intensive manual cell culture processes and analytics.  Analysis and decision points predominantly rely on generating large volumes of secreted protein, rate limited by cell doubling times and analytical sensitivities/development.  For every project, a wealth of data is generated.  However, a modular design framework has yet to leverage this data, unify processes, and transform current workflows.  Through the integration of machine learning frameworks, we are shifting from an established, traditional method of manual cell culture and analysis towards a “digital cell culture” paradigm.  Our strategy is two-fold: one is the implementation of nanofluidic technology to manipulate and collect data at a single-cell level, the second is utilization of machine learning to provide predictive power for cell line selection, deliver efficiency, and enable a scalable/transferable digital platform across Amgen.  Integrating machine learning into our workflow will ultimately enable precise decision making, drive improved performance, better understand/control quality attributes, and dematerialize workflows.

Picking winners: Predictive modeling for cell line selection

2019 Spring National Meeting

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.