Use of DOE Models for Clone Screening and Process Parameter Set Points for Scale Up (2026-EU-30MP-2859)

The biopharmaceutical industry's fast pace challenges its R&D teams to develop processes and deliver products in shorter timelines. To reach this goal, we are leveraging the use of high throughput (HT) devices during the process development phase. HT devices, such as ambr250 in upstream development and robocolumns in downstream development, are standard tools in operation. 

During upstream process development, selecting clones and defining the set points of process parameters are crucial steps for advancing quickly to the clinical phase. Various process parameters (temperature, biomass, perfusion rate, and production medium) significantly influence the quality of the drug substance. Traditional design of experiments approaches required a large number of experiments. To streamline this, we implemented a D-optimal factorial design using two ambrs.

In this presentation, we outline how to establish an effective design that offers strong predictive power. We will define a matrix of key performance indicators (KPIs) by leveraging prior knowledge and evaluating downstream impurity clearance capabilities. This KPI matrix allows us to select the best performing clone, as well as the process conditions needed for a stable perfusion process of this clone. This methodology enabled us to develop a statistical model validated across three scales, correlating all process parameter effects, their interaction effects, and quadratic effects on the KPIs. The outcomes of this work significantly decrease the number of experiments during upstream process development, while providing a valuable tool for scale-up predictions in the event of deviations and frontloading process characterization efforts.