Modeling mixed fermentation of gowé using selected Lactobacillus plantarum and Pichia kluyveri strains.

This paper describes a mixed fermentation model made by assembling block hosting models for the growth of lactic acid bacteria (Lactobacillus plantarum) and a yeast strain (Pichia kluyveri), metabolic production and the physical-chemical changes which occur during the fermentation of gowé. The growth model for P. kluyveri was developed on a synthetic medium following the gamma concept taking into account the effect of pH, temperature, concentrations in glucose, lactic acid and ethanol. Additional parameters for the previously defined L. plantarum growth model were also determined (glucose and ethanol concentrations). The model was validated in three different gowé processing conditions. Even if the model underestimates LAB growth, it explains what occurs in the product and enables in silico extrapolation to various fermentation conditions. The predicted hydrolysis rates of native and gelatinised starches showed that increasing malt content is not an efficient way to increase the sweetness of gowé in contrast to increasing the level of pre-cooking. The builing-block model developed in this study could be applied to many other fermented foods and particularly to non-alcoholic but acid and sweet cereal based beverages.