Multi-fold enhancement in sustainable production of biomass, lipids and biodiesel from oleaginous yeast: an artificial neural network-genetic algorithm approach

Abstract

Oleaginous yeasts have emerged as a much favored feedstock for sustainable production of biomass, lipids and biodiesel because of their higher growth rates, greater lipid contents and ease of cultivation as compared to plants or microalgae. In this study, the Plackett–Burman design was first implemented for screening the critical nutrients and then an artificial neural network modelling coupled with genetic algorithm (ANN-GA) technique was employed for enhancing the lipid content in Meyerozyma caribbica (earlier reported as Pichia guilliermondii) with greater biomass yield. Plackett–Burman screening experiments indicated that glycerol, ammonium chloride, magnesium sulphate and potassium dihydrogen phosphate were the most influential media components, whose concentrations were subsequently optimized by applying the ANN-GA technique. The optimized media, while doubling the biomass concentration, resulted in an enhanced lipid yield of 0.49 ± 0.02 g g⁻¹, which is approximately 2.5 fold the initial starting value as obtained in a 3.7 L fermenter. Based on gas chromatographic analysis of a fatty acid methyl ester (FAME) profile, the ratio of saturated to unsaturated fatty acids was found to be 44.5 : 55.9, which is considered as most favorable combination for biodiesel applications. The biodiesel properties also conformed to the ASTM D6751 and EN 14214 specifications, thereby making it a marketable clean and green biodiesel product. Thus, the present study showcases successful implication of an advanced media optimization strategy for multi-fold enhancement of biomass and lipid yields for sustainable production of biodiesel as a renewable fuel using Meyerozyma caribbica.