Many chemical processes are involved in the interactions of living cells with their environment; however, monitoring such processes often requires sophisticated analyzers. In this study, a sensing strategy based on imaging techniques has been developed to (i) enable cell discrimination based on their physical appearance such as size and shape and (ii) to build predictive models that relate the measured cell appearance to chemical parameters in their environment. Both goals aim at innovative and straightforward sensing strategies for analyzing cell–environment interactions. Image analyses offer several advantages such as the use of simpler, more robust sensors and the omission of extensive sample/sensor preparations. Imaging can analyze numerous cells and thus gains a culture representative insight rather than a potentially nonrepresentative single-cell response.
As a proof-of-principle application, different species of microalgae cells have been exposed to various nutrient conditions. Microalgae are known to sensitively adapt to changing nutrient conditions and could potentially become biological “probes” for chemical shifts in ecosystems. Because of considerable spreads of cell size and shapes within one class, size and shape distributions have been derived from visible images of cell cultures. It is shown that the novel image analyses are capable of discriminating different cell species based on their cell shapes and sizes. It is also demonstrated that in conjunction with the recently introduced, nonlinear multivariate “predictor surfaces”, the nutrient availability has a quantifiable impact on the cell size distributions. In this application, predictor surfaces are somewhat more precise than partial least squares. Copyright © 2012 John Wiley & Sons, Ltd.