Assessing Reproducibility Challenges in Clinical Metabolics

Current metabolomics research is expanding to be multiplatform and interdisciplinary. Metabolomic analysis has the potential to provide detailed insight into any biological system of interest to analyze the underlying molecular mechanisms of variables like environment, disease, and genetics. Human clinical metabolomics is a major research area that is often focused on finding biomarkers or panels of biomarkers for different diseases. However, results are difficult to compare across different laboratories since the field of metabolomics has yet to widely adopt best practices for sample preparation, data acquisition and data processing, results are difficult to compare across different laboratories. This thesis focuses on identifying the inconsistencies in current clinical metabolomics and then offering detailed examination on how experimental and processing parameters across the study can affect the detected metabolic signatures. The first portion of this thesis consists of a large-scale meta-analysis of cancer biomarker research to determine if there are consistent and true signals being detected for any specific cancers or broadly across all cancers. Results suggest that there is a high noise level and that metabolites need to be independently detected by several studies before they can be considered a true signal. The second half of this thesis analyzed experimental parameters used by NMR metabolomics studies to determine how variations in experimental settings and protocols affect the detection and measured concentration of metabolites. Outcomes of this portion of research have the potential to widely impact the field of metabolomics and help guide research labs towards adopting more standard practices and improving the field. Finally, this dissertation concludes with a study of in vitro fermentation seeking to understand the effects of both high and low carbohydrate conditions. Additionally, Chapter 7 is a brief communication that uses ion mobility mass spectrometry to separate near mass, multiply charged peptides and glycopeptides. Overall, this body of work has demonstrated that metabolomics is a valuable field but that it still has room to grow, adapt and improve. Advisor: Robert Powers