Media optimization for biosurfactant production by Rhodococcus erythropolis MTCC 2794: artificial intelligence versus...

Area of Science:

• Microbiology
• Biotechnology
• Bioprocess Engineering

Background:

• Biosurfactants are microbial-derived surface-active compounds with diverse applications.
• Optimizing production conditions is crucial for cost-effective biosurfactant yield.
• Rhodococcus erythropolis is a known producer of valuable biosurfactants.

Purpose of the Study:

• To optimize media components for enhanced biosurfactant production by Rhodococcus erythropolis MTCC 2794.
• To compare the efficacy of artificial neural network (ANN) coupled with genetic algorithm (GA) against response surface methodology (RSM) for media optimization.

Main Methods:

• Media optimization using ANN-GA and RSM, with sucrose, yeast extract, meat peptone, and toluene as variables.
• Development of ANN and RSM models to predict biosurfactant yield (% EI(24)).
• Comparative analysis of ANN-GA and RSM predictive accuracy using a separate experimental dataset.

Main Results:

• The ANN-GA model demonstrated higher accuracy and consistency, with a correlation coefficient of approximately 0.99 and average quadratic error of approximately 3%.
• RSM showed a higher average quadratic error of approximately 6%.
• ANN-GA optimized media resulted in a 3.5-fold increase in biosurfactant yield compared to unoptimized conditions.

Conclusions:

• ANN-GA is a superior and more reliable method for optimizing biosurfactant production compared to RSM.
• ANN-based models are effective for sensitivity analysis in bioprocess optimization.
• The study successfully achieved a significant enhancement in biosurfactant yield through optimized media formulation.