MSACL 2023 Abstract

Global Metabolomics and Lipidomics Platform for Ultra-High Performance Molecular Phenotype Analysis Shuang Zhao (1) and Liang Li (1)(2) (1) The Metabolomics Innovation Centre(TMIC), Edmonton, Alberta, Canada (2) Department of Chemistry, University of Alberta, Edmonton, Alberta, Canada Shuang Zhao, Ph.D. (Presenter) The Metabolomics Innovation Centre

Presenter Bio: My name is Shuang Zhao and I am currently on a research fellowship position in The Metabolomics Innovation Centre (TMIC) at University of Alberta, Canada. My research focuses on developing innovative LC-MS-based metabolomics and lipidomics solutions for various clinical applications.

Introduction
Comprehensive and quantitative characterization of molecular phenotype of biological samples is crucial for biological studies and health research. Dealing with detection and quantification of small molecules, metabolomics and lipidomics become emerging tools to study molecular phenotype in a time-sensitive and treatment-sensitive manner. Due to the significant differences in chemical and physical properties of metabolites and lipids, these two classes of compounds are analyzed separately. The accurate, reliable and meaningful molecular phenotype characterization relies on using proper and high performance metabolomics and lipidomics approach. Here we present a deep and global metabolomics and lipidomics platform for molecular phenotype analysis. The proposed technology was applied to different types of samples in proof-of-concept studies.

Methods
The metabolomics was performed using 4-channel chemical isotope labeling (CIL) LC-MS approach. Briefly, after metabolite extraction, samples were aliquots to four parts and derivatized with a pair of isotope reagents (i.e., 12C/13C-reagents) for each channel, followed by LC-MS analysis. Data processing was then preformed using a dedicated software IsoMS Pro.

The lipidomics was performed using in-depth global lipidomics approach. A set of rationally designed stable isotopic labeled internal standards (SIL-IS) was mixed with each sample, followed by lipid extraction. LC-MS and LC-MS/MS analyses were performed in positive and negative ion mode for each sample. All acquired data was processed by software LipidScreener.

The data from metabolomics and lipidomics can be interpreted either separately or combined together to achieve ultra-high coverage.
 
Preliminary Results
In the CIL LC-MS metabolomics, four submetabolomes were analyzed using isotopic reagents, i.e., amine/phenol, carboxyl, hydroxyl, and carbonyl submetabolome. With optimal design of the reagents, concomitant improvement in separation, detection and quantification can be achieved. To enable comprehensive identification, a three-tiered and dual-identification approach was employed: in tier 1 and 2, high confidence results can be generated using authentic standards or pathway-related metabolites, respectively; in tier 3, MS search was applied. With benefits from both labeling and comprehensive identifications, high performance analysis can be realized with much stronger ability for scientific discovery.

In the in-depth lipidomics, unique SIL-IS were created to improve the quantification of common lipid types. MS and MS/MS data were acquired to enhance both lipids coverage and identification. A three-tiered identification approach was developed. In tier 1 and 2, MS/MS data was used for positive identification with different score thresholds; in tier 3, MS data was used for putative match. With high coverage and better quantification using unique SIL-IS, high performance lipidomics can be accomplished.

Workflows of analyzing common samples have been developed. As an example, 18 human serum samples were analyzed using this technology. In metabolomics, 9179 ± 71 metabolites per sample were detected and relatively quantified. Among them, more than 1500 metabolites were identified with high confidence, covering over 100 metabolic pathways. In lipidomics, 8385 ± 266 lipids per sample were detected. Among them, more than 1000 lipids could be identified using RT, MS and MS/MS. Using this technique, a total of more than 2500 metabolites and lipids could be identified with high confidence and relatively quantified with high accuracy and precision, indicating an exceptionally high coverage and performance for molecular phenotype characterization. In the presentation, the analysis power of using integrated dataset of metabolomics and lipidomics for studying samples with different phenotypes will be shown.

Conclusions
A global metabolomics and lipidomics platform was developed for deep characterization of molecular phenotype.