Study On Computational Methods Of Chromatogram And Two Dimensional Electrophoresis Image

Computational analytical technology provides powerful tools for the pharmaceutical and biological research, now these computational methods have been widely used to process the data from chromatography, spectrometry and electrophoresis instruments. However, in the complicated mixture system present computational methods can't quantify the desired component accurately. In addition, in the protein spots detection, current computational methods are easily disturbed by the non-protein spots on two dimensional gel images, therefore the correct detecting rate of the protein spots can hardly be improved. So in this paper research work was carried out around the methodology study concerning the resolution of the overlapping peaks in chromatograms and protein spots detection on protein electrophoresis gel image using modem instrumental analysis, informatic skills and intelligent computational methods multidisciplinarily. The full research work includes the following parts:1. A new computational method for quantitative determining the desired component of complex pharmaceuticals is proposed. The Principle Component Analysis (PCA) is integrated with Generalized Rank Annihilation Factor Analysis (GRAFA) to process HPLC/DAD data and the traditional IND function is revised. This new method can be used to compute the desired component in the complicated mixture. In the paper this method is applied to compute the concentration of the component named notoginsenoside R| in the compound Chinese medicine, which is composed of Radix Salviae Miltiorrhizae and Radix Notoginseng. Good results indicate that the new method is more stable and accurate than GRAFA. Besides, the modification of IND function substantially enhances the accuracy of computation.2. A new method to detect protein spots from two dimensional electrophoresis gel image is proposed. Based on the color, area and shape differences between protein spots and non-protein spots, image-sharpening method, edge-detecting method and morphological feather extracting method were integrated to detect protein spots from two dimensional electrophoresis gel image. This new method can be used for detecting the protein spots from the gel image with strong disturbance. Both this method and PDQuest 7.2 were applied to detect the protein spots from a serum proteomic gel image. The result shows that the method proposed in this paper was more accurate and reliable than PDQuest 7.2 in detecting protein spots on gels with strong disturbance.The protein spot detection method proposed in this paper and two dimensional electrophoresis skill are used to analyze the serum proteome of healthy people and patients with benign and malignant breast tumor. GSTM5 is found highly expressed only in the patients' serum. By further analysis, it can be known that GSTM5 has much significance on the tumor prognosis and therapeutics.