Metabolite identification and databases

RALF WEBER: Metabolomics datasets collected applying mass spectrometry, including direct infusion mass spectrometry and liquid chromatography mass spectrometry, are composed of many hundreds of mass-to-charge ratio measurements arising from molecular ions, adduct ions, and naturally occurring isotopes. Metabolite annotation operates first by determining the molecular formulae from the accurate mass-to-charge ratio measured for each peak. A single peak can typically be assigned multiple molecular formulae, which can represent numerous naturally occurring structural isomers. Isomers have the same molecular formula, same number of different elements, such as carbon and oxygen, but these elements are constructed differently in 3D space. One set of isomers are leucine and isoleucine.

To be able to uniquely identify metabolites we must collect information on the metabolites structure, as this is unique for each metabolite. To acquire information on a molecule’s structure we can break the metabolite apart using fragmentation techniques inside a mass spectrometer. We usually call this tandem mass spectrometry when this process is performed in a single event. Fragmentation techniques impart energy to the metabolite which is present in a vacuum. Normally, energy is lost via molecule-molecule collisions at atmospheric pressure when energy is transferred from the energetic metabolite to many other molecules. However, in a vacuum this cannot occur, as there are only a very small number of gas molecules present.

So instead chemical bonds linking elements break as a way to disperse their energy. After a metabolite has been broken down into fragments, the mass spectrum is measured. This mass spectrum is unique for each metabolite, though fragmentation mass spectra can be similar for metabolites with similar chemical structures for a single fragmentation experiment. Information within a tandem mass spectrometry experiment is often not rich enough to differentiate between metabolites with the same mass and similar, but not identical, chemical structure. Therefore, the fragmentation process can be repeated, applying one or multiple fragment ions observed within a tandem mass spectrum. This will result in a multi-stage, or MS to the n, fragmentation tree.

The data collected from fragmentation mass spectrometry experiments are searched against a mass spectral library or database containing fragmentation mass spectra acquired for pure chemical standards. We can call this library searching or spectral matching and is an essential element of a fragmentation experiment. There are several public mass spectrometry libraries or databases available that can be employed for spectral matching, such as METLIN, Mass Bank, and mzCloud. Spectral matching algorithms make a comparison between experimental or reference mass spectra. Similarity scores are calculated for all comparisons. Ranking the similarity scores should result in the most correct assignment on top of the list. As discussed before, during this week, there are different levels of reporting confidence in metabolite identification and annotation.

Multi-stage fragmentation is often used as an additional analytical technique to improve the level and confidence of metabolite identification. However, the diverse nature of metabolites, the wide range of experimental parameters involved in a fragmentation experiment, and the mass spectral libraries that currently do not cover metabolite space make the collection and identification of metabolites a time consuming and challenging process.