Abstract
FTIR micro-spectral images of Caki-2 cells cytospun onto calcium fluoride (CaF2) slides were used to build a computational model in order to discriminate between the biochemical events of the continuous cell cycle during proliferation. Multivariate analysis and machine learning techniques such as PCA, PLSR and SVMs were used to highlight the chemical differences among the cell cycle phases and also to point out the need for removing the distortion of the spectra due to the morphology of the cells. Results showed cell cycle dependant scattering profiles that enabled the training of a SVM in order to recognise, with a relative high accuracy, each cell cycle phase purely with the scattering curve removed from the FTIR data after being subject to the RMieS-EMSC algorithm.