Model fit to experimental data

The model was compared with four different sets of anaerobic growth data on single and mixed sugars (Fig. 2.2). As a measure for the quality of model fit, coefficient of determination (also referred to as R2) is presented for each component of Figs. 2.2(a) to (d) (Table 2.3). R2 is defined as follows:

r2 = 1 _ SSn.; SSerr =X(yi, exp — yi, model )2/ SStot =X(yi, exp _ Vexp fz (5)

SStot i i

where yi exp, yi model, and yexp denote experimental data, their associated modeled value, and the mean of the observed data, respectively. R2 values are very high (i. e., over 0.9) for major components (such as glucose, xylose, biomass and ethanol). R2 values of minor components (such as glycerol and xylitol) are relatively low which is possibly due to the error introduced in data reading from literature graphs. Average R2 values are over 0.8 in all cases.

image099

Fig. 2.2. Comparison of model simulations with experimental data. Substrates: (a) glucose only, (b) xylose only, (c) and (d) mixed sugars. Symbols: □ glucose, О xylose, V ethanol, О cell dry weight, + glycerol, ^ xylitol, — simulations.

Fig. 2.2

(a)

(b)

(c)

(d)

Glucose

0.997

0.972

0.982

Xylose

0.974

0.985

0.990

Cell dry weight

0.988

0.929

Ethanol

0.936

0.926

0.956

0.957

Glycerol

0.788

0.857

0.501

0.480

Xylitol

0.828

0.838

0.729

Average

0.927

0.903

0.850

0.828

Table 2.3. Coefficient of determination (or R2) for individual components of Figs. 2.2(a) to (d).