In the development of noninvasive optical biopsy, normal tissues can be statistically differentiated from precancerous and cancerous tissues by analyzing their autofluorescence spectra. The observed cancer hallmarks in the spectra are manifestations of biochemical and morphological changes in tissue during cancerous transformation. For detection of colorectal cancers, it has been hypothesized that the major contributors to tissue fluorescence are three endogenous fluorophores – reduced nicotinamide adenine dinucleotide (NADH), flavin adenine dinucleotide (FAD) and collagen. Separating and identifying endogenous fluorophores in cells/tissues using capillary electrophoresis (CE) with laser–induced fluorescence (LIF) detection holds promise as a simple and fast method to analyze fluorophore compositions in tissues during the cancerous transformation. To this end, we have established the extraction and separation protocols for quantifying endogenous fluorophores in Chinese Hamster Ovary (CHO) cells, human colorectal adenocarcinoma cells (HT–29) and human normal colon cells (FHC). Flavin mononucleotide (FMN), FAD, NADH and nicotinamide adenine dinucleotide phosphate (NADPH) have been identified in the cell extracts by spiking them with standards and quantified by standard addition methods. The influence of cell densities and cell growth stages on fluorophore composition has been closely examined. Two–dimensional (2D) correlation coefficient mapping of electropherograms of HT–29 and FHC cell extracts reveals that the HT–29 cell extracts with higher cell density can be differentiated from FHC and HT–29 cell extracts with lower cell density, which is also demonstrated by the comparison of peak area ratios of NADH and NADPH. The electropherograms for 2D correlation analysis are pretreated by aligning their prominent peaks to account for peak shifting. A challenge in biological spectroscopy of cells and tissue is the identification of endogenous components that contribute to the overall complex spectra and the diagnostic signature. We propose 2D generalized correlation of CE–LIF electropherograms and fluorescence spectra in order to resolve the overlapped fluorescence spectra into their individual components. Separation of the endogenous fluorophores in normal and cancer cells by CE–LIF has provided us insight into fluorophore compositions and tools for classifications of cells. It has also prepared us for extraction and separation of tissues under different physiological conditions to assist cancer diagnosis.