Investigating the effect of media composition on growth and mAb production in CHO cells using a piecewise hybrid dFBA-PLS framework

In this study, we present a hybrid modeling framework that integrates piecewise Partial Least Squares (PLS) regression with Dynamic Flux Balance Analysis (dFBA) to simulate and optimize Chinese Hamster Ovary (CHO) cell fed-batch culture. Twenty-four Ambr15 experiments were conducted to systematically vary feed and inoculum compositions. Time-resolved metabolite, biomass, and Monoclonal antibodies (mAb) concentrations were collected and modeled. The hybrid model achieved high prediction accuracy (Normalized Mean Squared Error (NMSE) < 0.15 for most metabolites) and provided interpretable flux profiles. Multivariate analysis revealed consistent metabolic signatures tied to media formulation, where specific feed–inoculum combinations drove shifts in glycolysis, TCA cycle flux, and nitrogen metabolism. These insights demonstrate the model’s capacity to capture key metabolic adaptations and support data-driven media optimization in CHO cell culture. • A piecewise hybrid model developed and validated for CHO fed-batch culture. • Bioreactor experiments were performed with 24 different media blends. • A balanced production\consumption of key metabolites results in higher mAb titer. • Specific feed–inoculum blends drive shifts in glycolytic and TCA cycle fluxes. • Hybrid model results in good predictions for varying media composition (NMSE<0.15).