A multi-level meta-heuristic algorithm for the optimisation of antibody purification processes

Multi-product biopharmaceutical facilities need flexible process configurations that can adapt to products with diverse characteristics and impurity loads so as to avoid bottlenecks and delays, whilst meeting final product specifications and cost targets. In order to aid the design of such facilities, this work presents a meta-heuristic optimisation approach using genetic algorithms where different levels of decision are addressed (facility, product sequence and unit operation) and multiple process and business criteria are used to evaluate each alternative generated. This is applied to a case study on the production of therapeutic monoclonal antibodies (mAbs), with a focus on the optimal purification sequences and chromatography column sizing strategies to cope with different facility configurations of upstream and downstream trains and different product impurity loads. The industrial case study provides novel insights that allow the identification of the most cost-effective purification sequences and column sizing strategies that meet demand and purity targets for each product in the facility. Emphasis is placed also on providing methods to visualise the trade-offs in the set of optimal solutions with similar cost values so as to enhance the decision making process.

► A decisional tool to optimise biopharmaceutical purification was created.
► An industrial case study on the production of therapeutic antibodies was addressed.
► Chromatography sequences were optimised to meet cost and impurity removal targets.
► Genetic algorithms optimised column sizing strategies to meet time constraints.
► Graphical tools visualised user preferences for selecting alternative strategies.