MULTIVARIATE STATISTICAL-ANALYSIS APPLIED TO X-RAY-SPECTRA AND X-RAY MAPPING OF LIVER-CELL NUCLEI

Principal Components Analysis (PCA) and Factorial Analysis of Correspondence (FAG), two Multivariate Statistical Analyses (MSA), were applied to the analysis of X-ray data. MSA are descriptive methods which graphically display the correlations and anticorrelations between a large number of elements. Series of X-ray spectral data and X-ray maps obtained from rat liver were analyzed with reference to the diffusible elements Na, Mg, Cl, K and Ca and also P and S. By using an in situ precipitation method, the pyroantimonate method, it was found that the free, precipitable cations Na+, Mg2+, and Ca2+ are, in the nuclei, mainly distributed throughout the nucleoplasm. Images obtained from FAC allow those areas rich in nucleic acids to be displayed as areas with a strong anticorrelation between P and Sb. In cryoprocessed tissues, the evaluated wet mass-fraction of diffusible elements corresponds to physiological values of total amounts (free and complexed). PCA makes it possible to graphically display the correlation between P and K in chromatin and nucleolus, the correlation between K, Cl and S in cytoplasm and nucleoplasm and the observation of two populations of nuclei according to different Na, Mg and K concentrations. Factorial images obtained from FAC allow those areas rich in nucleic acids to be displayed as areas with a strong correlation between P and K. Similarly those areas rich in proteins are displayed as areas with a strong correlation between S and K.