Metabolomics-based methods for cancer biomarker discovery: application to esophageal and bladder cancers

Jian Zhang, Purdue University

Abstract

Metabolomics is a growing field in systems biology and offers a powerful and promising approach to identify biomarkers associated with numerous diseases including cancer, diabetes, and inborn errors of metabolism. Nuclear magnetic resonance (NMR) spectroscopy and mass spectrometry (MS) are the two most commonly used analytical methods in metabolomics. The two methods are complementary; while NMR is highly quantitative and reproducible, MS is highly sensitive. Utilization of both NMR and MS methods leads to the analysis of over 1000 small molecules (molecular weight <1000 >Da). In this dissertation, a combined NMR and MS-based metabolomics approach was used to detect human esophageal diseases such as esophageal cancer, Barrett's esophagus (BE) and high grade dysplasia (HGD), and canine bladder cancer. Analyses of the complex NMR and MS data using advanced multivariate statistical methods provide enormous possibilities for metabolomics research and discovering disease biomarkers. Newly-developed analytical techniques involving LC-TOF MS and NMR that are proven to be powerful methods for metabolic profiling were used in the biomarker discovery. A number of highly sensitive biomarkers were identified using these advanced techniques and various multivariate statistical methods. Partial least square (PLS) and partial least square-discriminant analysis (PLS-DA) after univariate Student's t-test based feature (metabolite) selection create robust mathematical models to detect significant differences between disease and healthy subjects arising from perturbations caused by disease processes. Key metabolites were identified from the statistical results and then validated as biomarker candidates. Based on the identified biomarker candidates, intrinsic disease-related mechanisms were evaluated and suggested for further studies. In particular, emerging technologies in metabolomics discussed in this thesis are shown to be effective, and open a number of potential avenues for further development and clinical applications.

Degree

Ph.D.

Advisors

Raftery, Purdue University.

Subject Area

Analytical chemistry|Biochemistry|Oncology

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS