Artificial Intelligence Applications for Producing Glycosylated Biopharmaceutical Drug Modalities

Biopharmaceutical drug modalities derived from animal cell culture are inherently complex and heterogeneous by nature. Among the most complex of these biologic entities are therapeutic glycoproteins and glycosylated viral vectors employed in Advanced Therapeutic Medicinal Products (ATMPs). The specific profile of the glycans attached to these modalities constitutes a relevant critical quality attribute (CQA) of such biologics. Data on the structure and composition of these product-associated glycans are the most difficult to obtain at high quality and throughput. In combination with the wealth of data obtained routinely for the tightly controlled process variables during upstream production, such detailed glycan data are much needed to enable artificial intelligence tools to derive information and knowledge for the optimization of product glycosylation during biomanufacturing. In this contribution, recent developments in this field will be discussed.