Identifying Variables Influencing Traditional Food Solid-State Fermentation by Statistical Modeling

Abstract

Solid-state fermentation is widely used in traditional food production, but most of the complex processes involved were designed and are carried out without a scientific basis. Often, mathematical models can be established to describe mass and heat transfer with the assistance of chemical engineering tools. However, due to the complex nature of solid-state fermentation, mathematical models alone cannot explain the many dynamic changes that occur during these processes. For example, it is hard to identify the most important variables influencing product yield and quality fluctuations. Here, using solid-state fermentation of Chinese liquor as a case study, we established statistical models to correlate the final liquor yield with available industrial data, including the starting content of starch, water and acid; starting temperature; and substrate temperature profiles throughout the process. Models based on starting concentrations and temperature profiles gave unsatisfactory yield predictions. Although the most obvious factor is the starting month, ambient temperature is unlikely to be the direct driver of differences. A lactic-acid-inhibition model indicates that lactic acid from lactic acid bacteria is likely the reason for the reduction in yield between April and December. Further integrated study strategies are necessary to confirm the most crucial variables from both microbiological and engineering perspectives. Our findings can facilitate better understanding and improvement of complex solid-state fermentations.