Automated bioprocess systems rely on accurate online measurements of cellular growth. In the present study, six sensors representing five different technologies, namely near-infrared absorbance, external non-invasive light scattering, back scattering reflectance, dielectric spectroscopy, and exhaust gas analysis were evaluated in online monitoring of the growth of the microbial cell factory Corynebacterium glutamicum. The signal outputs were monitored during the stationary growth phase of the cells under different conditions of agitation and aeration, conditions that pose potential interference with the in-situ signals. The random noise of the measurements was filtered using a Gaussian filter and the time derivatives were computed to obtain information on the microbial growth rate. Surprisingly, the signals of most probes under fed-batch cultivations of C. glutamicum were highly correlated with the stirring, but were mostly unaffected by the airflow, despite the increase in air bubbles that can cause measurement interference. Taken together, we show that an online estimation of the growth of C. glutamicum based on the different sensor signals was possible with linear and multiple linear regression, correcting for the effect of stirring, and an estimation of growth rate was made accurate with an implementation of a Gaussian filter.
