Artificial Intelligence as a Combinatorial Optimization Strategy for Cellulase Production by Trichoderma stromaticum AM7 Using Peach-Palm Waste Under Solid-State Fermentation

Abstract

The use of agro-industrial waste as substrate to produce enzymes has been widely studied in order to adjust the high productions of these biocatalysts with a reduction of the final cost in an ecologically correct process in the industry. This is the first work to use peach-palm (Bactris gasipaes Kunth.) waste for the production of carboxymethylcellulases (CMCase) by the Trichoderma stromaticum AM7 under solid-state fermentation (SSF) using the artificial neural network and genetic algorithm (ANN-GA) to optimize the cellulase production. ANN-AG was used to determine the optimal influence of nitrogen source concentration (2.46%), time (12 days), and temperature (26 °C) on the cellulase production with 98% of efficiency for algorithm prediction of endoglucanase activity. The predicted and experimental values of the CMCase activity in U/g of dry initial substrate (gds) were 122.9 and 120.0, respectively. After all optimization processes, the variation of these parameters allowed a final increase of 31.58-fold when compared to the initial fermentation. Moreover, the treatment of peach-palm waste with T. stromaticum AM7 endoglucanase had an effect on the degradation of the fibers analyzed by scanning electron microscopy and the saccharification of the waste by releasing reducing sugar (807.80 mg/g) for 8 h of incubation. These results are very promising to waste valorization and the biotechnological and industrial applications of cellulase of the T. stromaticum AM7 to food processes and biofuel production.