Abstract
This study evaluated two kinetic models for the enzymatic synthesis of amoxicillin catalyzed by penicillin G acylase, using the Markov chain Monte Carlo (MCMC) method for estimating the process's kinetic parameters. The first model, based on Michaelis-Menten kinetics, and the second, founded on reaction and equilibrium constants, were optimized using a single initial condition and subsequently validated under 12 distinct experimental conditions. Sensitivity analysis enabled the identification of the most sensitive parameters for each model, while the selection of the model that best fits the experimental measurements was based on Bayesian metrics and the relative mean squared error. The model based on reaction and equilibrium constants demonstrated superior predictive capability, exhibiting a 18.40% error after optimization compared to the 25.79% observed in the Michaelis-Menten model. These results underscore the efficacy of integrating mathematical modeling, Bayesian statistics, and sensitivity analysis in predicting amoxicillin production under different experimental conditions.