The high-throughput ability and sensitivity of mass spectrometry make it the most popular platform for a metabolomics study. Thus, identifying small molecules from mass spectra plays a central role in metabolomics. There are many computational techniques for mass spectra raw data processing, including feature detection, peak alignment, and mass spectral deconvolution. This dissertation focuses on the conversion between spectral and structural information. It is a spontaneous thought to identify compound structure from mass spectra by searching the query spectra against a reference library with similarity score. However, this approach is limited by the availability of reference spectra and standard compounds. To bridge the gap, different computation tools based on fragmentation trees are developed to help annotate spectra. The other thought is to generate in-silico spectra from the structural information. Several compound databases, such as PubChem, KEGG, HMDB and CHEBI can be the source of structural information. Machine learning and heuristic approaches are trained from spectral knowledge to generate in-silico spectra. However, all those approaches cannot predict beyond the known data. As a first-principles method, Quantum chemistry modeling is only based on the rules of quantum mechanics and can help us explore the unknown space of metabolites. However, the quantum chemical simulation on molecules over 600 Da is too expensive to get accurate results. Chemical ionization with methane reagent gas can help identify the molecular ion species and help the unknown metabolite identification.This dissertation describes computational studies using molecular dynamics on in-silico mass spectra of small molecule generation. Chapter 1 provides an overview of applications of quantum chemistry to generate in-silico mass spectra. Chapter 2 showcases the performance of quantum chemistry simulations on small molecules and probes the parameter space to find potential improvements to the existing method. The conformational flexibility effect is also explored and has no correlation with prediction accuracy. Chapter 3 describes a new workflow to include chemical derivation, especially silylation in the quantum chemistry calculation. Different compound classes including organic acid, alcohol, amide, amine and thiols are simulated and compared against experimental mass spectra. The molecular dynamics trajectories are also investigated to find missed fragmentations from rearrangements. Chapter 4 provides a new algorithm that introduces excited state calculation into the molecular dynamics prediction of mass spectra. The new algorithm can predict more fragmentation reactions that are missing in previous studies and as a result, the mass spectral similarity scores are increased, and simulation is more accurate. Chapter 4 also discusses the limitations of molecular dynamics simulation time and the lack of rearrangement reactions. Chapter 5 provides another routine from the experimental side to help unknown metabolites identification with methane chemical ionization and quadrupole-time of flight mass spectrometry.

Metabolomics is a powerful tool in the field of natural compounds research, offering insights into the complex metabolic processes occurring within living organisms. In the context of natural compounds, metabolomics enables the identification and characterization of the diverse array of chemical entities produced by plants, animals, and microorganisms. This approach provides a holistic view of the dynamic nature of biological systems, unraveling the intricate interplay between genes, proteins, and metabolites, and facilitating the discovery and development of novel bioactive compounds for various applications, including medicine, agriculture, and biotechnology. This thesis explores the application of cutting-edge technologies in the identification and characterization of natural compounds, shedding light on their diverse chemical structures and potential biological activities. In Chapter one, a straightforward and potent method is presented for uncovering glycosides in food, specifically focusing on chocolate and tea, which are known to be rich in glycosides. Building upon the observation that glycosides and their corresponding aglycones fragment in almost identical patterns, a library of 3,027 standards was subjected to in-silico conjugation using thirteen different sugar combinations. This approach illuminated the vast realm of unknown glycosides present in food, revealing that virtually any compound can be conjugated at any point in time. Chapter two presents how a metabolomics approach enabled a combinatorial screen of 109 Family 1 Glycosyltransferases with 578 natural compounds. The experiments involved employing cell lysate as the source of enzymes, and substrates were grouped in sets of 40 to enable efficient high-throughput screening. The samples were processed in 96-well plates and subsequently analyzed through the utilization of LC-MS/MS technology. A comprehensive analysis resulted in the annotation of over 250 compounds, suggesting their potential substrate activity with at least one enzyme. This innovative approach not only facilitated the identification of enzymes capable of glycosylating aliphatic compounds but also revealed that Histidine is not a crucial residue for UGT1 activity. Chapter three discusses the effect of precursor ion intensity in the identification confidence of metabolites. We injected 200 metabolites in 13 different concentrations, by collecting negatives and positive mode data in 3 different collision energy, which resulted 15,600 spectra with different spectral quality. By comparing high- and low-quality spectra, we found that normalized entropy can be used for labeling low-quality spectra. Furthermore, our findings indicate that implementing denoising techniques on spectra can significantly enhance the reliability of compound identification, offering a novel dimension of data management for scientists in the field of mass spectrometry. Finally, Chapter four evaluates a direction that the field of metabolomics has been striving to embrace: the integration of targeted and untargeted metabolomics in a single LC-MSMS experiment. Previous endeavors have demonstrated success in conducting both targeted and untargeted experiments within a single injection. However, these efforts have been constrained by the limitations of instruments equipped with only a single detector, which restricts the ability to employ both approaches within a specified timeframe. This concept is made possible by the utilization of a Tribrid mass spectrometer, which incorporates two detectors concurrently. In our study, we effectively merged a Parallel Reaction Monitoring (PRM) analysis employing an Ion Trap with an untargeted analysis utilizing an Orbitrap in the IQ-X Tribrid Mass Spectrometer. By employing this method, we successfully identified low-abundance compounds that would have otherwise been overlooked by conventional untargeted analysis techniques.

Metabolism occurs in a spatial-temporal fashion and represent the product of cellular functions. Mass spectrometry-based metabolomics enables the analysis of hundreds to thousands of analytes that can be used to study biological processes in the context of disease. Here, we leverage novel mass spectrometry approaches- both high-resolution mass spectrometry coupled to liquid chromatography, as well as matrix assisted laser desorption ionization mass spectrometry imaging- to exam neurometabolic related questions.Chapter one characterizes the biochemical, behavioral, and cytological in the brain in long-evans rats following neonatal, postnatal day six (P6), exposure to DMSO, the widely used vehicle (VEH) solvent. First, neurometabolic changes were profiled 24 hours after exposure in four distinct brain regions- the cortex, hippocampus, basal ganglia, and cerebellum- using hydrophilic interaction liquid chromatography. Second, behavioral tests were performed between P21-40 to investigate the chronic alterations to behavior following short term, early exposure. Lastly, immunofluorescence and immunohistochemistry were used to assess cytological changes in microglia, astrocytes, and neurons at P40. These findings show that short term exposure of DMSO, regardless of dose, at neonatal stages alters key regulatory metabolites and alters neurochemistry, results in chronic hypoactive behavior and decreased social habits, and in dose dependent increase in microglia and astrocytes. Chapter two investigates the role of p73a1, a p73 C-terminal isoform, in regulating lipid metabolism. Characterization of this novel protein isoform was described by Dr. Kyra Laubach, and cell lines were generated using CRIPSR to remove E12 in cancer cell lines. A multi-omic approach was integrated with molecular biology techniques to identify lipid classes and lipid metabolism-associated genes that were altered by the expression of p73a1. Furthermore, the tumor suppressive function of p73a1 was found to be mediated in part through lipid metabolism-associated genes. Through these findings, a previously unidentified p73 target was established, in addition to determining a role for p73a1 in lipid metabolism. Chapter three provides a framework and application for spatial metabolomics using matrix assisted laser desorption ionization- mass spectrometry imaging (MALDI-MSI). I demonstrate that spatial metabolomics enables data driven segmentation that provides unique clusters that resemble and map to anatomical structures. Next, I show that lipids in the brain are highly organized, and these lipids can be identified in tissue through MS/MS experiments. Additionally, I perform manual segmentation to extract ion intensities for regions of interest and show that MSI enables histopathological analysis of small molecules. Finally, I apply this workflow to a case study of litter-matched matched rats, one that expresses transgenes that induce an Alzheimer-like pathology, and its wild-type, healthy control litter-mate. I show that several phosphatidylcholine (PC) lipids are altered in the entire brain and specific regions of interest. Other PC lipids are changed only in clear ROIs when comparing WT and AD brains but are not changed in the whole brain. This workflow is a critical step necessary to enable histopathological analysis using metabolomics.

Metabolomics is an analytical approach for systematic profiling of metabolites in biofluids, cells and tissues. Gas chromatography - mass spectrometry (GC-MS) and liquid chromatography - mass spectrometry (LC-MS) are most popular techniques in metabolomics due to their high-throughput and high sensitivity. Common procedures in metabolomics include study design, sample collection, sample extraction, data acquisition, data normalization, statistical analysis, and biological interpretation. All procedures before data normalization may introduce analytical variations as a hurdle for statistical performance. With the purpose of improving quality assurance in metabolomics for human disease studies, my dissertation work started with applying internal standards and external quality controls for quality assurance in untargeted metabolomics as showed in chapter one. A new tool called Systematical Error Removal using Denoising Autoencoder (SERDA) was developed and the median relative standard deviations (RSD) of the training QC samples was reduced to 4.6% RSD after normalized by SERDA. Then, I performed untargeted metabolomics assays to explore the lipidomic alterations and risk prediction of Type 2 Diabetes (T2D) in the cohort of American Indians from the Strong Heart Family Study (SHFS). Multivariate analysis of lipidomics identified distinct lipidomic signatures that can differentiate high- from low-risk groups. Higher baseline level of 33 lipid species, including triacylglycerols, diacylglycerols, phosphoethanolamines, and phosphocholines, was significantly associated with increased risk of T2D at 5-year follow-up. Because studies in humans usually can only denote risk factors but not biochemical mechanisms, I continued my work on animal and cell studies in chapter three and four. In chapter three I showcase how metabolomics data can be integrated with phenotype information in mouse models to decipher the link between gene functions and metabolism and downstream diseases. Finally, I compared how different technologies can be used to measure isotope flux analyses in bacterial cell samples to detail specific differences in metabolic pathways under anaerobic or aerobic conditions. The comparison results revealed that all three instruments can provide similar biological conclusions with each instrument possessing their own advantages.