Multi-way signal processing for multicolor fluorescent biosensing
Abstract
In this paper, we investigate the possibility of modelling the concomitant response of multiple bacterial genes to environmental factors by using fluores-
cent whole cell biosensors. An experimental procedure is proposed to obtain three-way data sets by synchronous fluorescence spectroscopy (SFS). A Candecomp/Parafac algorithm (CP) is then proposed to separate simultaneously all the fluorescence signals of reporter proteins and to interpret the response of the corresponding promoters. The joined estimation of several gene responses is the main original point of this blind identification procedure. The method proves to be more powerful than the traditional principal components analysis (PCA) or the singular value decomposition (SVD) and provides a promising multi-way strategy for in vivo monitoring of gene expression in system biology studies.