Mechanistic and data-driven modeling of protein glycosylation

Modulation of glycosylation in therapeutic proteins is a critical aspect to their development and production. The levels of various glycan moieties greatly impact the therapeutic protein's overall efficacy and safety. As such, controlling the glycan levels and understanding potential levers that impact them is highly desirable. Various computational tools exist to understand these levers and quantify their impact on this critical quality attribute (CQA). Here we present a review on recent advances of these computational tools, how these advances further our understanding of the glycosylation pathway, and their potential applications. We focus on both mechanistic models for N-linked glycosylation, including the vesicular and maturation model, for predicting glycosylation profiles and providing insights into the glycosylation pathway itself. We also discuss data-driven models for predicting glycosylation profiles and identifying process levers for glycosylation.

Automated summary

Modulation of glycosylation is crucial in the development of therapeutic proteins, as different glycan moieties can significantly affect efficacy and safety. Computational tools, including mechanistic models and data-driven models, are used to understand and quantify the impact of glycosylation on therapeutic proteins. These tools help researchers gain insights into the glycosylation pathway and identify process levers for glycosylation.