| Metabonomics, the scientific research of endogenous metabolic responses of living systems to stimuli, serves not only as a source of qualitative but also quantitative data of metabolites essential for the description of the metabolic cycle. As an application-driven science, metabonomics has been applied to the curatorial area broadly. For the questions of metabonomic research, the quantitative research strategy and integrated research strategy were proposed, and the data processing was carried out with a new try. These strategies and methods were used in the study of major diseases, which resulted in achieving good results.For the inaccurate quantitation of nowaday metabonomics, quantitative research strategy was applied to the metabonomic research. And our own research methods were proposed. Firstly, aimed at purine and pyrimidine, an analytical platform for simultaneous quantification of 21 related metabolites was established using HPLC-UV-MS/MS. Then the platform was applied to the research on diabetic nephropathy. With the research, seven potential biomarkers were found out, which were uric acid, xanthine, inosine, adenosine, cytosine, cytidine and thymidine. And the possible mechanism of the disease was speculated. Afterwards, an analytical platform for simultaneous quantification of 16 metabolites involved in folic acid, homocysteine and glutathione metabolism was established using HPLC-MS/MS. Then the platform was applied to the research on the avoidance of neural tube defects (NTDs) with nutriment. With the research, four potential biomarkers were found out, which were Hcy, 5-MT, GSH and GluCys. The results illuminated that the nutriment was effective refered to the data of NTDs patients. And the possible mechanism of the nutriment was speculated. Through the complete research, the problem of poor quantitation was solved by using the quantitative research strategy. So the results of metabonomic research were more reliable and acceptable.For the problem of the unilateral content and complex object, integrated research strategy was applied to the metabonomic research. And our own research methods were proposed. Firstly, we integrated quantitative metabonomics and metabolic fingerprinting to look for comprehensive biomarkers, which make the prediction accuracy rate of identification for disease stage reaching to 96%. Then the metabonomics was integrated with clinical research. With the combination of metabolic biomarkers and clinical biochemical parameters, the prediction accuracy for disease was increased. And the disease mechanism was speculated with the combination. Finally, a research mode was put forward for finding out metabolic pathway related to disease with the integration of genomics and metabonomics. With the research mode, a very important metabolic pathway related to diabetic nephropathy was found out, which is linoleic acid metabolism. Integrated strategy strengthen the links among the various studies, so that the research for the disease became deeper and more thorough. For the poor capacity of solving multiclass and prediction, artificial intelligence technologies were applied to the data processing. With the combination of fuzzy logic and artificial neural network, a new method was established for finding out biomarkers, which was then applied to the research of diabetic nephropathy. For the data of quantitative metabonomics, we found out four potential biomarkers, which were uric acid, cytosine, xanthine and thymidine. Compared to seven biomarkers found out using student-t test, they had similar prediction accuracy. For the data of metabolic fingerprinting, the variables was reduced from 4000 to 4, while both of them had good prediction accuracy, which is higher than 0.92. From the research results, we can see that the established method can achieve multiclass classification with a good prediction accuracy. In addition, the variables can be reduced, therefore the potential biomarkers can be found out with the artificial intelligence technologies. |
PDF Full Text Request