(582z) Investigation of the Impact of Bioprocess Variables on Performance using Multivariate Analysis

Multivariate analysis of process data can be used to discover new relationships between process variables or to optimize performance of a production cell line. These studies can be used for optimization of late stage manufacturing process or to refine existing platform processes. In this study, we relied on historical data from a variety of small to medium scale cell culture projects and products in order to identify correlations between inputs like lactate production or off-line pH measurements and outputs like specific productivity or titer. The scope of the study includes several Chinese hamster ovary (CHO) cells, producing monoclonal antibodies (mAbs) and fusion proteins. Data from multiple scales was included in the analysis, from the 2 liter scale through the 200 liter scale, as well as data from different stages of the process development timeline. The process data was primarily obtained from off-line samples, including off-line pH, metabolites, and product titer, with additional parameters like mean-adjusted lactate production derived from these values to reveal additional correlations between process variables. The investigation uncovered interesting connections between metabolites like specific lactate productivity and specific productivity. Future multivariate analysis will expand the scope of analysis and investigate the opportunities for improvement at production scale. This study demonstrates the power of multivariate analysis of process parameters through the conclusions that can be drawn from previously existing data, and produces additional value at little cost.