Metabolomics refers to studies of the entire complement of metabolites present in an organism or sample. Methods to reliably detect and identify all metabolites in a sample are lacking however. LC-MS is a preferred approach for analysis because of the versatility, sensitivity, and chemical information content. Current LC-MS methodology enables tens of thousands of "features", i.e. signals at a given retention time and m/z, to be detected from complex mixtures of metabolites. However, identification of these features remains a challenge. Also, it is likely that many more metabolites exist than can be detected. We desribe two approaches to improve compound detection and identification. In the first, we investigate high resolution LC based on small particles and long columns operated at 35 kpsi to enhance the number of compounds detected and identified. We find substantial increases in identificatoin with long columns generating peak capacity of 600 or greater.  In the second, we test the hypothesis that low S/N of MS/MS spectra prevent good quality matches to database for many detected features. Strategies to increase sample loading and other steps show that we can increase identified compounds to over 1500 from a single sample. We further show that another subset of detectable features have high quality MS/MS spectra but are not present in databases for identification based on matching.

Chromatography-centric approaches to improving compound identification in metabolomics

ACS Fall 2021

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.