Optimization of fungi co-fermentation for improving anthraquinone contents and antioxidant activity using artificial neural networks

The fermentation products of edible fungi are rich in anthraquinones and have a variety of activities, including the antioxidant activity. Because of the large number of combinations, it is very difficult to obtain the optimal multi-strains co-fermentation to improve the yield of anthraquinone. In the present study, an intelligent model based on artificial neural networks (ANNs) using backpropagation (BP) and radial basis function (RBF) algorithms was developed and validated to predict the anthraquinone contents in 136 two fungi and 680 three fungi co-fermented products. After experimental validation of the anthraquinone contents, the mean absolute error and the mean bias error of the results from RBF ANN were lower than those from BP ANN. The results indicated that the anthraquinone contents in A. bisporus, C. comatus and H. erinaceus co-fermentation product was the highest (2.11%). Furthermore, this co-fermentation product showed strong antioxidant activity.