Rice Bran or Apple Pomace? Comparative Data Analysis of Astaxanthin Bioproduction

Abstract

Modeling and optimization of high value-added astaxanthin pigment bioproduction statistically by Sporidiobolus salmonicolor ATCC 24259 from two substantial wastes, rice bran (RB) and apple pomace (AP) was aimed in this study. The experimental data was obtained at constant inoculum rate (2%) and particle size (0.85 mm) for both wastes by conducting 17 runs, which were generated by Box-Behnken design. 33.41 µg astaxanthin gRB- and 77.31 µg astaxanthin gAP- were produced as the maximum amount at the end of fermentation period, 10 days. Apple pomace was concluded as the optimized waste for the production of astaxanthin based upon the highest yield. Predicted response results of response surface methodology (RSM) and radial basis function-neural network (RBF-NN) were compared in order to evaluate the accuracy of two methodologies on non-linear behavior of the astaxanthin bioproduction. RBF-NN became prominent with its well-suited to apple pomace fermentation system by resulting in quite low 0.8495, root mean square error (RMSE), 0.3349, mean absolute error (MAE), and 0.9985, correlation coefficient (CC) as best measures of a model performance. 

Keywords

• apple pomace; rice bran; astaxanthin; radial basis function-neural network