Alchemy of Fungi and AI: Response Surface Methodology and Artificial Neural Network‐Based Optimization of Echinocandin B Production in <i>Aspergillus nidulans</i>

Echinocandin B (ECB) is the precursor of a first-line antifungal, anidulafungin, widely used to treat systemic and invasive fungal infections in nosocomial and community-acquired settings. This potent antifungal is naturally synthesized in Aspergillus nidulans in trace amounts, which can be improved by optimizing growth and physiological conditions. The current study is focused on optimizing the fermentation medium by employing statistical and artificial neural network (ANN)-based model to improve ECB activity in fermentation broth. In the present study, the most significant parameters for ECB activity (i.e., molasses, dextrose, casein, and pH) were identified through the Plackett-Burman design and were further optimized using different statistical models based on central composite design experiments. Process optimization with a reduced quadratic (RQ) model and ANN model (architecture: 4-6-2-1) suggested a 3.64- and 3.03-fold increase in ECB activity, respectively. However, prediction from the RQ model (R²: 0.93) could be unreliable when compared to the ANN model (R²: 0.99), effectively capturing the complex relationships. The study concludes that the ANN-based predicted model displayed more accuracy and provided optimum levels of the analyzed factors for a 3-fold increase in ECB activity.