Files
Abstract
Metabolomics is the study of small molecules in a biological system. It has aided in research on characterizing health conditions, screening out important pathways, and disentangling associations between metabolites and pathways. Because of the diversity and complexity of metabolites, data processing and analysis are important steps for metabolomics research. Nuclear magnetic resonance (NMR) spectroscopy is one of the most powerful analytical approaches to metabolomics. However, the difference in sample properties such as pH may cause great variations in chemical shifts of some NMR resonances. Although efforts have been made to control this variation, this phenomenon still challenges NMR-based metabolomics research. In addition, metabolomics data can be integrated with other types of data for systematic analysis,but techniques for data integration have not been well explored yet. Therefore, there is a demand for advancing techniques for NMR-based metabolomic analysis.
In this dissertation, I demonstrated two novel computational tools I developed for NMR-based untargeted metabolomics. These include a data integration technique for analyzing relationships between NMR-measured metabolomics, MS-measured glycomics, and two-dimensional flow-cytometric data; and a spectral alignment algorithm, named pHIT, for processing NMR peaks with high chemical shift variations.
By applying these tools to the study of C. elegans development, and pregnancy under normal and challenged conditions, I discovered new biological phenomena. I identified development-associated metabolite-glycan correlations in C. elegans. I analyzed urine metabolome from pregnant mothers with virus infection and found disturbed metabolites and pathways. I then used a sheep model to further investigate fetal and neonatal metabolic change with maternal stressed conditions as well as their metabolic transitions after birth. I identified preterm-birth-associated metabolic change with maternal chronic cortisol treatment and altered metabolites and pathways after birth in neonates.
My work is expected to contribute to NMR-based metabolomics analysis, as well as the understanding of metabolic change in animal development and gestation.
In this dissertation, I demonstrated two novel computational tools I developed for NMR-based untargeted metabolomics. These include a data integration technique for analyzing relationships between NMR-measured metabolomics, MS-measured glycomics, and two-dimensional flow-cytometric data; and a spectral alignment algorithm, named pHIT, for processing NMR peaks with high chemical shift variations.
By applying these tools to the study of C. elegans development, and pregnancy under normal and challenged conditions, I discovered new biological phenomena. I identified development-associated metabolite-glycan correlations in C. elegans. I analyzed urine metabolome from pregnant mothers with virus infection and found disturbed metabolites and pathways. I then used a sheep model to further investigate fetal and neonatal metabolic change with maternal stressed conditions as well as their metabolic transitions after birth. I identified preterm-birth-associated metabolic change with maternal chronic cortisol treatment and altered metabolites and pathways after birth in neonates.
My work is expected to contribute to NMR-based metabolomics analysis, as well as the understanding of metabolic change in animal development and gestation.