Computer Assisted Information Processing On Pharmaceutical Chromatographic Analysis

Modern pharmaceutical analysis is confronted with comprehensive information derived from diversified samples with the development of advanced analytical techniques and instruments. Analytical emphasis has been transferred to unearth the effective result from the giant data pool, leading to the emergence and gradual blossom of chemometrics. Recent progress in algorithm and computer-assisted technology accelerates the evolution of various chemometrics branches, i.e., condition optimization, multi-variant calibration, digital simulation, pattern recognition, data mining and so on. Our paper focuses mainly on the optimization and fingerprinting in the chromatographic analysis of traditional Chinese medicines (TCM).Improved simplex method is studied to optimize the capillary zone electrophoresis(CZE) condition of two Nitrocamprothecinum isomers. A chromatographic response function named CRS is modified on CZE to guide the correct direction of optimization. Compared to the traditional single factor optimization, improved simplex method simplifies the optimization with lower number of experimental steps. The method is further employed to optimize the extracting condition of Rhizoma Anemarrhenae. Inappropriate extraction parameters and interference from determination errors result in the inefficacy of optimization, however, indicating the significance of a suitable response and searching area. Corresponding software was programmed to assist optimizations with lower demand for analysis experience of operators.An gradient HPLC method is established and fully validated before determining epimedin A, C, icariin, sagittatoside A and baohuoside I in the extract of Epimedium Brevicornum crude drugs. Thereby, chromatographic fingerprints of batches of crude drugs are collected and come out to be useful in their quality control. Further study is carried out to evaluate the possibility of TCM fingerprinting based on the HPLC-MS technique. Since mass spectrometry is one of the accomplished approaches in elucidating the chemical structure of an unknown component even with low concentration, the HPLC-MS-based fingerprinting of TCM is more competent than the conventional one. HPLC-MS data should be supplemented to the HPLC fingerprint information for a secondary peak matching.BP artificial neural network is then constructed by MATLAB 5.3 to recognize the fingerprints of Epimedium Brevicornum by the pattern recognition model based on the fingerprints of TCM. The model discriminates 5 categories of crude Epimedium Brevicornum from each other successfully.Inspired by the success of blended wine, attempts are made for the first time to simulate the trueborn crude Epimedium Brevicornum by blending the crude Epimedium Brevicornum of various sources. The improved simplex method is adopted again to get the best mixing ratio of different kinds of non-genuine crude drugs in order to improve the similarity based on the fingerprints respectively. Optimized blending of crude drugs may be efficient in making full use of relatively inadequate TCM resources.In conclusion, condition optimization, pattern recognition and other computer-assisted information processing methods in pharmaceutical analysis would play more and more important role in the quality control or even quality improvement of TCM.