AI-Assisted Response Surface Methodology for Growth Optimization and Industrial Applicability Evaluation of the Diatom Gedaniella flavovirens GFTA21

Abstract

Although AI-mediated approaches provide promising support for bioengineering using training datasets, their application in microalgal research remains limited. In this study, ChatGPT-4.0, an easily accessible AI model, was employed to optimize culture conditions and evaluate the industrial potential of the isolated diatom Gedaniella flavovirens. Culture optimization was conducted using response surface methodology, in which pH, temperature, and salinity were selected as independent variables. ChatGPT assisted in determining the design and suggested a face-centered central composite design. The optimal conditions for biomass production were determined to be pH 8.30, 23 degrees C, and 34.24 psu. Analysis of variance revealed significant quadratic effects (p < 0.05), indicating curvature in the response surface. Fatty acid profiling showed high levels of palmitoleic acid, palmitic acid, and eicosapentaenoic acid. Pigment analysis further indicated a high abundance of fucoxanthin, diadinoxanthin, and diatoxanthin. Based on the analyzed compounds, ChatGPT suggested potential applications of the algal strain across various industrial sectors. The most relevant application was identified as aquafeed, as the strain contains metabolites known to enhance pigmentation, growth, and immune responses in aquaculture species. Overall, this study demonstrates ChatGPT-mediated bioengineering as a practical strategy for optimizing culture conditions and evaluating the industrial potential of novel microalgal strains.