Response surface methodology and artificial neural network based media optimization for pullulan production in...

Area of Science:

• Biotechnology
• Microbial Fermentation
• Biopolymer Production

Background:

• Pullulan, a microbial polysaccharide, has diverse industrial applications.
• Optimizing fermentation conditions is crucial for efficient pullulan yield.
• Aureobasidium pullulans is a key microorganism for pullulan biosynthesis.

Purpose of the Study:

• To screen and optimize carbon and nitrogen sources for enhanced pullulan production.
• To identify significant factors influencing pullulan yield using experimental designs.
• To develop and validate a predictive model for pullulan fermentation.

Main Methods:

• Screening of various carbon (sucrose, fructose, glucose) and nitrogen sources ((NH4)2SO4, urea, NaNO3).
• Plackett-Burman and Central Composite designs for factor identification and optimization.
• Development of a hyperparameter-optimized artificial neural network (ANN) model for prediction.

Main Results:

• Sucrose and NaNO3 were identified as optimal carbon and nitrogen sources, yielding 9.33 g L-1 pullulan.
• KH2PO4, NaCl, and sucrose were significant factors identified by Plackett-Burman design.
• The ANN model achieved high predictive accuracy (R2: 0.96) and predicted a yield of 31.9 g L-1, closely matching the observed 30.17 g L-1.

Conclusions:

• Optimized fermentation conditions using sucrose and NaNO3 significantly enhance pullulan production.
• The artificial neural network model provides a reliable and accurate method for predicting pullulan yield.
• This study demonstrates a robust approach for optimizing microbial polysaccharide fermentation.