Anomalies of deconvolution via discrete Fourier transform: a case study on assessing transport at live cell preparations

Abstract

Deconvolution is one of the most frequently used mathematical operations in the evaluation of data in many branches of science and engineering. Analytical chemistry is no exception: deconvolution is part of routine methodologies in major instrumental techniques such as many modalities of spectroscopy and chromatography. The need for proper deconvolution, however, is often overlooked in the emerging field of bio-analytical chemistry as applied to live tissue in vivo, or cells in vitro. It is in this context that some of the major problems inherent to deconvolution, and to one of its most commonly used methodologies, deconvolution via discrete Fourier transform (DFT), are explored in this work. To illustrate the nature and severity of the difficulties involved and the resulting errors and potential for misinterpretation, the problem of assessing cellular processes with different microsensors is chosen as an example here. In particular, cancer multidrug resistance (MDR) modulated doxorubicin (DOX) efflux from drug sensitive cells derived from Chinese hamster ovary cancer, and their MDR subline, are discussed. DOX concentration monitored with a carbon fiber microelectrode at close proximity to a monolayer of these cells can be perceived as the result of convolution of cellular efflux with extracellular diffusion. Deconvolution via DFT to reconstruct the efflux process, however, leads to patterns that are very far from realistic. The observations and conclusions made using this example are relevant to other areas of analytical chemistry, and science and engineering in general that use deconvolution methods in processing experimental results of many types.