By using liquid chromatography/mass spectrometry (LC/MS), thousands of peaks can be detected in a metabolite extract from a typical biological sample. The unbiased and comprehensive profiling of these peaks is known as untargeted metabolomics. In contrast to targeted approaches which focus on only a subset of these molecules, untargeted metabolomics is global in scope and presents an unprecedented opportunity to interrogate previously unexplored metabolic pathways at the systems level. Despite the global scope of the untargeted approach, the overwhelming majority of metabolomic publications to date have exclusively applied targeted methods. A critical barrier that has prevented the widespread and large-scale application of untargeted metabolomics is the time and expertise required for data interpretation, specifically to establish metabolite identification. To directly address this barrier, the proposed work will develop a new untargeted metabolomic workflow in which the metabolite identification process is automated. The automated platform will accelerate the identification of large numbers of metabolites by requiring significantly less time and expertise. To support the automated platform, this proposal will develop new software which will link what is currently the most widely used metabolomic software (XCMS) with the largest metabolite database (METLIN). Importantly, XCMS and METLIN have a longstanding history of being freely available and the proposed software will therefore be highly adoptable by the general scientific community. The developed software will automatically perform two major functions: (i) relative quantitation and (ii) database searching for identification on the basis ofthe accurate mass of the intact compound as well as its tandem MS spectra. Other functionalities that will guide the non-specialist in identifying unknown compounds will also be incorporated, such as molecular classification and pathway mapping. Additionally, the software will provide a tool to perform meta-analysis across independent studies. In the latter context, the proposed work will enable the ultimate large-scale analysis by facilitating the comparison of untargeted metabolomic data from multiple labs.
Despite the great potential of metabolic screening for diagnostics and pathological insight, global studies of metabolites has been limited by the time and expertise required for data interpretation. We propose developing an accelerated workflow based on a software program that will automate data interpretation such that global studies of metabolites can be performed by non-experts as part of routine biomedical analyses.
Showing the most recent 10 out of 29 publications
