Improvement of Predictive Model to Reduce Impurity Presence Based on Raw Material Combination Pre-Selection

Abstract

Biotechnology manufacture is composed of two main sections, Active Pharmaceutical Ingredient (API) and Drug Product (DP). The DS manufacturing most common steps are cell culture or fermentation (DSI), recovery and purification. The purification process purpose is to reduce impurities to acceptable levels. Depending on the purification raw materials, there will be different interactions with the DSI to be purified that could affect clearance capability. Multivariate analysis includes the effect of the considered variables and the response of interest. Therefore, it can be used to take into consideration combinations DSI solutions and purification raw materials available in inventory to predict the results of the impurity of interest. In this project, the improvement of a multivariate predictive model was performed by gathering recent manufacturing scale data and defining a more representative data set to improve the accuracy of the predictions. Upon implementation of the updated predictive model, several consecutive lots resulted in impurity results below the acceptable criteria. Key Terms - Bioprocessing, Biotechnology, Raw Materials, Prediction, Predictive Model, Purification