Computers & Chemical Engineering, Vol.27, No.1, 45-54, 2003
Investigating the fermentation dynamics structure of recombinant yeast YPB-G
The system of recombinant Saccharomyces cerevisiae growing in a batch reactor is highly nonlinear, and its fermentation characteristics are of particular interest. In this paper, a new algorithmic approach is proposed to determine the generic form of the system under study. Experiments were conducted at different initial starch concentrations, and a variety of data concerning biotic and abiotic phases were collected. Making use of various interdisciplinary nonlinear data analysis tools, the phase space defined by the measured quantities was reduced using the phase space dimension and the mutual information and correlation coefficients between measured quantities. This choice of reduction was validated through the eigenvalues of the covariance matrix of the data set by a Karhunen-Loeve type analysis, and the functional dependencies of possible models describing the system behavior were obtained. It is shown that, a generic form reached in a reduced phase space eases future experimental effort by 33%, and decreases the computational complexity by about 75%, while increasing the parameter reliability considerably. The data processing and analysis framework proposed may form a basis for further work in the structured modeling of biological systems, recombinant yeast cells in particular.