Optimization of the production of bacterial cellulose using multivariable linear regression analysis

The biosynthesis of bacterial cellulose by Acetobacter xylinum was optimized by numerically finding the maximum of an arbitrarily chosen second order polynomial model function of several variables (describing the dependence of the cellulose production on the concentrations of the medium components), using multivariable linear regression analysis. The chosen function appeared to describe the analyzed correlation sufficiently well. Consequently, three to six stages of optimization made the determination of the optimum medium compositions possible for 16 days of fermentation at 30 degrees C in a medium based on fructose (wt%: fructose, 3.68; yeast extract, 5.02; (NH4)(2)NO3, 0.001; KH2PO4, 0.3; MgSO4 x 7 H2O, 0.05; resulting in a cellulose production equal to 0.505 wt.% - namely 5.6 times higher than before the optimization) and for 7 days of fermentations at 30 degrees C in a medium based on sucrose and ethanol (wt.%: sucrose, 5.0; ethanol, 1.36; yeast extract, 1.27; (NH4)(2)SO4, 0.5; KH2PO4, 0.3; MgSO4 x 7 H2O, 0.05; resulting in a cellulose production equal to 0.251 wt.% - namely 1.5 times higher than before the optimization).