三种典型复杂生物体系代谢组学的NMR研究
发布时间：2018-07-28 13:12
【摘要】：代谢组学是一门通过分析复杂生物体系中小分子化合物的整体变化来探索生命过程本质的学科。NMR因其具有样品处理简单、谱编辑手段丰富、检测无偏向性、重复性好等优点,一直都是代谢组学研究最常用的一种检测技术。代谢组学研究的复杂生物体系包括生物体液、组织和器官等,这些体系含有大量各种各样的代谢物,而所有的生物代谢物又都含有质子。因此,基于1H NMR技术的代谢组学分析方法是研究复杂生物体系强有力的工具。基于NMR的代谢组学研究中,我们得到的原始数据通常是一组NMR图谱。使用化学计量学模型深度分析这些NMR图谱数据之前,我们首先需要对NMR图谱进行数据降维,以方便后续的统计分析。目前,最常用的NMR图谱降维方法是分段积分法,该方法将原始图谱分为若干个宽度为2 Hz左右的小区间,然后对每个区间积分得到积分谱。分段积分法的优点是高效、易于自动化,特别适合大规模样品研究；缺点是得到的数据变量多、图谱分辨率降低,为后期特征变量的归属埋下隐患。当样品数较少时,我们还可以采用分峰拟合法对NMR图谱进行降维。分峰拟合法利用各种代谢物的标准图谱对实验谱进行拟合得到代谢物的浓度数据,该方法的优点是得到的数据变量少,利于后续统计分析;缺点是对代谢物的图谱数据库依赖性强、不易实现自动化,因此仅适合样品数较少的初期研究。本论文第二章选取C57BL/6J大鼠建立转移黑色素瘤动物模型,并获取大鼠的脾脏和胃部组织为研究对象。在这一章中,我们首先选取大鼠的脾脏组织数据,分析和对比了分段积分和分峰拟合两种NMR图谱降维方法对主成分分析(Principal Component Analysis, PCA)模型和正交偏最小二乘分析(Orthogonal Proj ection to Latent Structures, OPLS)模型的影响。然后,我们分析了转移黑色素瘤在大鼠脾脏和胃部组织引起的代谢表型变化。研究结果表明,实验组和对照组样品具有明显的区分,转移黑色素瘤可能严重干扰了组织细胞中的多个代谢通路。本论文第三章采用与第二章相同的研究策略,选取C57BL/6J大鼠建立动物模型,分析了吸烟对正常体重和肥胖大鼠肺部组织代谢表型的影响。OPLS模型分析表明,不论大鼠是否肥胖,吸烟都会显著降低组织中腺苷衍生物ATP、ADP和AMP的含量,并且,肌苷和尿苷的含量会明显的升高。同时,我们还发现,仅吸烟并不会明显改变大鼠肺部组织中GPC/PC的比值,而肥胖这个单因素则足以显著提高GPC/PC比值,并且吸烟会增强这一变化趋势。本论文第二章和第三章研究的是不同组别样品间的差异,而忽略了组内样品的个体差异。组内样品的个体差异同样很重要,它可以反映不同个体对同一外部刺激的不同响应程度。因此,本论文第四章对样品的个体差异做了一些初步探讨。本章,我们选取布洛芬(IBP)为模型药物,研究了IBP与血浆相互作用的个体差异。本章内容是对本课题组先前一项研究工作内容的补充和延伸。在本章中,我们重新定义了相互作用指数(interaction index, Idist)和多样性指数(diversity index, Idiv)来表征血浆组分的个体差异及其对IBP-血浆相互作用的影响。重新定义的Idist能更多地提取NMR图谱中的信息,重新定义的Idiv具有唯一性和确定性,并且,Idist与Idiv之间表现出极为显著的线性相关性(p2.2e-16)。同时,我们分析了Idist与NMR图谱、Idist与临床数据以及临床数据与NMR图谱之间的关联。研究发现,IBP与血浆相互作用的个体化差异主要体现在血浆中小分子代谢物与脂蛋白结合强弱以及IBP与脂蛋白结合强弱的个体化差异。同时我们还发现,血浆中游离乳酸含量越高,或者总甘油三酯、白蛋白或胆碱浓度越低,对IBP越敏感；并且,体重较轻或者肥胖指数BMI较小者对IBP较敏感。本课题招募的这批志愿者中,年龄越大的人其血浆中葡萄糖、糖蛋白的浓度越大,缬氨酸、络氨酸、肌酐和脂蛋白颗粒中含胆碱磷脂的浓度越低。我们还可以通过OPLS模型的相关系数编码的负载图挖掘出NMR谱中与目标代谢物处于相同或者相近代谢通路的其他代谢物。
[Abstract]:Metabonomics is a subject to explore the essence of life process by analyzing the overall changes of small and medium molecular compounds in complex biological systems..NMR has the advantages of simple sample processing, rich spectrum editing, unbiased detection and good reproducibility. It has always been a most commonly used detection technique in metabolomics research. The complex biological systems include biological fluids, tissues and organs, which contain a large variety of metabolites, and all biological metabolites contain protons. Therefore, the metabonomics analysis method based on 1H NMR technology is a powerful tool for the study of complex biological systems. In the study of NMR based metabolomics, we have to The original data is usually a set of NMR atlas. Before using the chemometrics model to analyze these NMR data, we first need to reduce the data of the NMR atlas to facilitate the subsequent statistical analysis. At present, the most commonly used NMR map reduction method is a piecewise integral method, which divides the original map into a number of 2 widths. The advantages of the piecewise integral method are high efficiency, easy to automate, and are especially suitable for large-scale sample research. The disadvantage is that the obtained data are more variable, the resolution of the atlas is lower, and the later feature variables are buried. When the number of samples is less, we can also use the peaks. The NMR map is reduced by the fitting method. The peak fitting method uses the standard atlas of various metabolites to fit the experimental spectrum to get the concentration data of the metabolite. The advantage of this method is that the obtained data are less variable and benefit to the subsequent statistical analysis; the disadvantage is that the data base of the metabolite's Atlas data base is strong and not easy to automate, so it is only suitable. In the second chapter of this paper, C57BL/6J rats were selected to establish the animal model of metastatic melanoma and to obtain the spleen and stomach tissues of the rats as the research object. In this chapter, we first selected the spleen tissue data of rats, and analyzed and compared the subsection integral and the peak fitting of the two methods to reduce the dimension of the NMR map. Effects of Principal Component Analysis (PCA) model and orthogonal partial least square analysis (Orthogonal Proj ection to Latent Structures, OPLS). Then, we analyzed the metabolic phenotypic changes caused by metastatic melanoma in the spleen and stomach tissues of rats. The results showed that the experimental group and the control group were samples. The third chapter of this paper uses the same research strategy as the second chapter to establish the animal model of C57BL/6J rats, and analyzes the effect of smoking on the metabolic phenotype of lung tissue in normal weight and obese rats by.OPLS model analysis. No matter whether rats were obese, smoking significantly reduced the levels of adenosine derivatives ATP, ADP and AMP, and the content of inosine and uridine increased significantly. We also found that smoking alone did not significantly change the ratio of GPC/PC in the lung tissue of rats, and obesity was sufficient to significantly increase the GPC/PC ratio, and The second and third chapters of this paper are to study the differences between different groups of samples and ignore the individual differences in the samples in the group. The individual differences in the samples in the group are also important, which can reflect the different responses of different individuals to the same external stimuli. Therefore, the fourth chapter of this paper is on the sample. In this chapter, we selected Bloven (IBP) as a model drug and studied the individual differences in the interaction between IBP and plasma. The content of this chapter is a supplement and extension of the previous research work of this group. In this chapter, we redefined the interaction index (Idist) and The diversity index (diversity index, Idiv) was used to characterize the individual differences in plasma components and their effects on IBP- plasma interactions. The redefined Idist could extract more information from the NMR map, the redefined Idiv has uniqueness and certainty, and there is a very significant linear correlation between Idist and Idiv (p2.2e-16). At the same time, we analyzed the association between Idist and NMR atlas, Idist and clinical data as well as clinical data and NMR atlas. The study found that the individual differences in the interaction between IBP and plasma were mainly reflected in the individual differences in the combination of small and medium molecular metabolites with lipoproteins and the strong and weak combination of IBP and lipoprotein. Now, the higher the content of free lactic acid in the plasma, or the total triglyceride, the lower the concentration of albumin or choline, the more sensitive to IBP; and the lighter weight or the less obese index BMI is more sensitive to IBP. The older the volunteers recruited in this project, the greater the concentration of glucose and glycoprotein in the plasma, valine, complexon, and muscle. The lower the concentration of choline phosphatidylcholine in the anhydride and lipoprotein particles, we can also excavate other metabolites in the NMR spectrum that are in the same or similar metabolic pathways through the load map encoded by the correlation coefficient of the OPLS model.
【学位授予单位】：中国科学院研究生院(武汉物理与数学研究所)
【学位级别】：博士
【学位授予年份】：2016
【分类号】：Q591
，
本文编号：2150257
[Abstract]:Metabonomics is a subject to explore the essence of life process by analyzing the overall changes of small and medium molecular compounds in complex biological systems..NMR has the advantages of simple sample processing, rich spectrum editing, unbiased detection and good reproducibility. It has always been a most commonly used detection technique in metabolomics research. The complex biological systems include biological fluids, tissues and organs, which contain a large variety of metabolites, and all biological metabolites contain protons. Therefore, the metabonomics analysis method based on 1H NMR technology is a powerful tool for the study of complex biological systems. In the study of NMR based metabolomics, we have to The original data is usually a set of NMR atlas. Before using the chemometrics model to analyze these NMR data, we first need to reduce the data of the NMR atlas to facilitate the subsequent statistical analysis. At present, the most commonly used NMR map reduction method is a piecewise integral method, which divides the original map into a number of 2 widths. The advantages of the piecewise integral method are high efficiency, easy to automate, and are especially suitable for large-scale sample research. The disadvantage is that the obtained data are more variable, the resolution of the atlas is lower, and the later feature variables are buried. When the number of samples is less, we can also use the peaks. The NMR map is reduced by the fitting method. The peak fitting method uses the standard atlas of various metabolites to fit the experimental spectrum to get the concentration data of the metabolite. The advantage of this method is that the obtained data are less variable and benefit to the subsequent statistical analysis; the disadvantage is that the data base of the metabolite's Atlas data base is strong and not easy to automate, so it is only suitable. In the second chapter of this paper, C57BL/6J rats were selected to establish the animal model of metastatic melanoma and to obtain the spleen and stomach tissues of the rats as the research object. In this chapter, we first selected the spleen tissue data of rats, and analyzed and compared the subsection integral and the peak fitting of the two methods to reduce the dimension of the NMR map. Effects of Principal Component Analysis (PCA) model and orthogonal partial least square analysis (Orthogonal Proj ection to Latent Structures, OPLS). Then, we analyzed the metabolic phenotypic changes caused by metastatic melanoma in the spleen and stomach tissues of rats. The results showed that the experimental group and the control group were samples. The third chapter of this paper uses the same research strategy as the second chapter to establish the animal model of C57BL/6J rats, and analyzes the effect of smoking on the metabolic phenotype of lung tissue in normal weight and obese rats by.OPLS model analysis. No matter whether rats were obese, smoking significantly reduced the levels of adenosine derivatives ATP, ADP and AMP, and the content of inosine and uridine increased significantly. We also found that smoking alone did not significantly change the ratio of GPC/PC in the lung tissue of rats, and obesity was sufficient to significantly increase the GPC/PC ratio, and The second and third chapters of this paper are to study the differences between different groups of samples and ignore the individual differences in the samples in the group. The individual differences in the samples in the group are also important, which can reflect the different responses of different individuals to the same external stimuli. Therefore, the fourth chapter of this paper is on the sample. In this chapter, we selected Bloven (IBP) as a model drug and studied the individual differences in the interaction between IBP and plasma. The content of this chapter is a supplement and extension of the previous research work of this group. In this chapter, we redefined the interaction index (Idist) and The diversity index (diversity index, Idiv) was used to characterize the individual differences in plasma components and their effects on IBP- plasma interactions. The redefined Idist could extract more information from the NMR map, the redefined Idiv has uniqueness and certainty, and there is a very significant linear correlation between Idist and Idiv (p2.2e-16). At the same time, we analyzed the association between Idist and NMR atlas, Idist and clinical data as well as clinical data and NMR atlas. The study found that the individual differences in the interaction between IBP and plasma were mainly reflected in the individual differences in the combination of small and medium molecular metabolites with lipoproteins and the strong and weak combination of IBP and lipoprotein. Now, the higher the content of free lactic acid in the plasma, or the total triglyceride, the lower the concentration of albumin or choline, the more sensitive to IBP; and the lighter weight or the less obese index BMI is more sensitive to IBP. The older the volunteers recruited in this project, the greater the concentration of glucose and glycoprotein in the plasma, valine, complexon, and muscle. The lower the concentration of choline phosphatidylcholine in the anhydride and lipoprotein particles, we can also excavate other metabolites in the NMR spectrum that are in the same or similar metabolic pathways through the load map encoded by the correlation coefficient of the OPLS model.
【学位授予单位】：中国科学院研究生院(武汉物理与数学研究所)
【学位级别】：博士
【学位授予年份】：2016
【分类号】：Q591
，
本文编号：2150257
本文链接：https://www.wllwen.com/shoufeilunwen/jckxbs/2150257.html
