Representative examples of three cell lines (Fig. cells were acquired using an upright Confocal Raman microscope (alpha300RA, WITec GmbH, Germany) with a 100 oil immersion objective (NA 1.4) (Carl Zeiss, Germany). The sample was excited with a linear polarized (0) coherent compass sapphire green laser ex = 532 nm (WITec, Germany). The scattered Raman signal was detected with an optic multifiber (50 nm diameter) CNX-2006 directed to a spectrometer UHTS 300 (WITec, CNX-2006 Germany) (600 g mm?1 grating) and finally to the CCD camera (DU401 BV, Andor, North Ireland). Control Four (WITec, Germany) acquisition software was used for the Raman imaging set up. Spectra were recorded in 1 m X/Y steps for all samples with the exception of cells used for viable cell analysis, where spectra were acquired every 500 nm. The laser power was set to 30 mW and an integration time of 0.5 s was chosen to ensure fast mapping and to avoid cell damage. 2.3. Spectral pre\processing and data analysis All spectra were subjected to cosmic ray removal using a cosmic ray detection algorithm (filter size of two spectral pixels with a sensitivity indicated by a dynamic factor of eight) in the software Witec Project Plus 4.0. 2.3.1. Cluster analysis of cell averages: CHO\K1 and protein producers The Raman spectra of the image scans of the host cell line CHO\K1 and protein producers CHO\K1\hDAO and CHO\S\Humira were averaged using a mask filter in order to use only the areas comprising the cells. The average spectra of each cell were baseline corrected CNX-2006 and derived (second derivative, 17 smoothing points) prior to cluster analysis (wave number region 403C3750 cm?1), based on the Euclidean distance and follows Ward’s minimum variance algorithm: the distance between neighbors is given by heterogeneity (OPUS software, Bruker Optic GmbH, Germany). Spectra gathered in the same cluster were used to calculate the average spectra shown in Fig. ?Fig.1B1B and their second derivative in Fig. ?Fig.11C. Open in a separate window Figure 1 Raman microscopy spectra separate CHO host and producer cell lines. (A) Cluster analysis based on the spectral region 403C3750 cm?1 of the host line CHO\K1 (= 6 cells) and the two protein producers CHO\K1\hDAO (= 9) and CHO\S\Humira (= 5). All individual cells of each cell line were separated into different clusters indicating sufficient differences in their spectral characteristics. (B) Average spectra in the measured range of each cell line. (C) Second derivative spectra used for cluster analysis (baseline corrected and second derivative with 17 smoothing points). For (B) and (C) the spectra are stacked on the y\axis to allow for clear discrimination. 2.3.2. Cluster analysis of a Raman image: Cell and ER Raman images (20 20 m2, 400 spectra) of individual cells of a total of seven cell lines (CHO\K1, CHO\K1\hDAO, CHO\S\Humira, CHO\S, CHO\S/4F11, CHO\K11D9 and CHO\K14F10) were background subtracted (using a polynomial function of degree 3) and k\means cluster analyzed with the following conditions: four clusters, spectral mask 400C1800 cm?1 (fingerprint region), Manhattan normalization mode (area under the spectral mask equal to 1), no data reduction and no pre\transformation mode (Witec Project Plus 4.0, Witec, Germany). By this the hyperspectral image is processed to four images presenting the Tcfec chemically most different four areas (clusters) and the corresponding average spectrum for each distinguished area. One cluster was assigned to ER due to higher intensity of the protein bands, one to background (which was discarded) and the remaining two clusters were more related and merged as rest of the cell (RC). The average spectra of ER and RC of each cell, from the clustering, were used in a principal component analysis (PCA) based on the second derivative of those spectra (Savitzky\Golay Algorithm, 17 smoothing points) and the fingerprint region of 400C1800 cm?1 (Unscrambler X 10.3, CAMO Software While, USA). 2.4. PDI immunofluorescence staining Cells were fixed for 10 min with 4% formaldehyde\remedy.