Advances in analytical instrumentation are a key factor that enabled the development of metabolomics as a systems biology research tool. Continued advances in instrumentation have increased the number of detectable features in metabolomics data, but the proportion of unidentified features in a typical untargeted metabolomics study has remained high; this challenge continues to limit interpretation of potentially biologically informative data. Secondly, most attempts to identify features in metabolomics data are performed in an ad- hoc manner, and are usually undertaken only when biological data suggests differential abundance of an unknown feature between sample groups, and even then, identification is only pursued on a few of the highest- priority targets. While this approach superficially reduces analyst burden by limiting identification to only those features of biological interest, its actual effect is to increase the overall burden of compound identification, since these efforts are not carried out in a systematic manner, and their results are rarely indexed in major databases. Finally, MS/MS spectra are not routinely acquired for all target metabolites, and there exists no universal, cross-laboratory workflow for routine compound identification, so each laboratory is effectively required to set up its own extensive library of authentic standards to aid in compound identification. This is a costly proposition, which results in variable identification of even well-known metabolites. The Experimental Core of the Michigan Compound Identification Development Core (MCIDC) will help address these challenges by carrying out the following Specific Aims. First, we will attempt to identify recurrent unknown features in untargeted metabolomics data using a systematic, prioritized workflow. Secondly, we will develop and implement novel and cutting-edge analytical techniques and apply them to identification of unknown features in metabolomics data from biomedically-relevant samples. These techniques will include sample pre-fractionation and off-line multidimensional liquid chromatography, ultra-high-pressure separations, chemical derivatization and in-vivo stable isotope labeling, ion mobility mass spectrometry, and high-resolution NMR analysis. Finally, we will generate a high-quality metabolite library from the unknown features we identify in addition to previously characterized knowns, which will include a range of empirical details to aid in future compound identification efforts.
