It is hypothesized that cells with stem cell-like properties may be influential in carcinogenesis, possessing the ability to self-renew, produce differentiated daughter cells and resist environmental or therapeutic injury. This has led to a surge in interest in identifying and characterizing the tumour initiating or cancer stem cell (CSC) with the aim of discovering novel diagnostic and prognostic markers and of understanding the basic biology with the ultimate aim of generating new therapeutic approaches and biomarkers. However, a major hurdle to this process has been the lack of a truly specific cancer stem cell biomarker allied to the rarity of these cells. This has led to problems in characterising these CSCs by traditional '-omic' techniques. Using a renal carcinoma cell line model, we show that synchrotron radiation-Fourier transform infrared (SR-FTIR) spectroscopy is a suitable tool to measure discrete differences in the biochemistry of small numbers of single-cells. Using the chemometric techniques of Principal Component and Linear Discriminant Analysis (PCA and LDA) for multivariate reduction, biochemical differences between the cells from different sub-populations were evaluated. Results found lipid and phosphodiester vibrations to be particularly good discriminating markers in the spectra of these stem-like cells, relative to the more differentiated, proliferating cells that make up the majority of the cell population.