A signal-processing method known as spectral correlative chromatography (SCC) for two-dimensional data obtained from hyphenated chromatography is developed and applied to chemical chromatographic fingerprint data sets of herbal medicine under specific experimental conditions. The method can judge the presence or absence of a spectral correlative peak among the spectrochromatograms. A local least squares regression model (LLS) is constructed in a piecewise manner to correct the shifts of retention time of some peaks of interest in the chromatograms of various test samples. The results compare favorably with those obtained by a two-point calibrated algorithm. It is shown that performing SCC and LLS on the piecewise clusters of various chromatographic fingerprints is more helpful in practice in revealing their common nature and for characterizing the chemical constituents. This approach holds great potential for facilitating quality control of herbal medicines.

Key WordsGas ch romatog rag hy -mass spectrometry Chemometrics Su bwindow factor analysis Rehmannia glutinosa Libosch Scrophulariaceae Summary An evolving approach is described for analyzing the acidic components of a petroleum ether extract of Radix Rehmanniae Preparata. The latter was extracted, separated and derivatized. The subsequent separation of the individual components was characterized by GC-MS. The Iwo-way data were resolved into a pure chromatogram and a mass spectrum of each chemical component. The reliability and accuracy of the qualitative results were greatly improved and the degree of the chromatographic separation could be enhanced to a certain extent. This enabled the analysis of complicated practical systems by hyphenated instruments and these advanced methods.

No abstract