Plasma Metabolomics Research Of Coronary Heart Disease Based On GC-MS

Coronary heart disease (CHD) is one of the major causes of morbidity and mortality all over the world, afflicting millions of people every year. The risk factors of CHD are very complex. The researches about the cause and mechanism of CHD have great significance in early diagnosis and treatment of CHD. As a new discipline of systems biology, metabolomics has a unique advantage in the areas of diagnosis and treatment of disease, drug efficacy and toxicity evaluation, gene function, etc. A comprehensive study on the plasma metabolic profiles of healthy controls and CHD patients was done with the help of gas chromatography-mass spectrometry (GC-MS), effective derivatization protocol and chemometric methods such as multivariate resolution and pattern recognition in this thesis. The contents in this paper mainly includes several respects as followings:(I) GC-MS technology coupled with universal derivatization protocols were used to analyze the metabolic profiles of plasma from40cases of healthy controls and94cases of CHD patients. With the help of heuristic evolving latent projections (HELP), overlapping peaks of acquired data were effectively handled.46endogenous metabolites were identified and quantified.(II) Principal component analysis (PCA) and partial least squares linear discriminant analysis (PLS-LDA) were applied for the discriminant analysis between healthy controls and CHD patients. A PLS-LDA model was basically established, which got a good result for recognition and prediction. In addition, a newly proposed method that can be used to screen variables-Subwindow Permutation Analysis (SPA) was employed to screen variables, and several important metabolites contributing to the classification were screened out for potential biomarkers.(III) PCA and the Monte Carlo tree (MCTree) approach were applied for the discriminant analysis between CHD patients with different severity according to whether they had stenosis of one, two or three of the coronary arteries, and the separation trend for the three groups was significantly evident. In addition, the MCTree approach was employed to pick out the important variables for potential biomarkers, which may play important roles in the pathological changes of CHD. All the metabolites that selected as biomarkers were briefly concluded biologically. All these may provide useful information for early diagnosis, prediction and pathogenesis of CHD.