Abstract | Metabolite identification is still the major bottleneck in non-targeted metabolomics. Different levels of identifications have been proposed by the Metabolomics Society Identification Task group (Sumner et al 2007, plus paper with refined levels). The highest level of identification can be only achieved by comparison with an authentic standard using two independent properties such as retention times and MS/MS spectra. Having reference standards for all possible metabolites in a single laboratory is nearly impossible and not feasible. Several public repositories storing MS/MS spectra have been created, also covering various instrumentations including MS analyzers (e.g., Q-ToF, IT or ICR) (MassBank, Metlin etc.). MS/MS spectra represent only one part of an accurate identification and can be ambiguous in case of isomeric compounds, wrong collision energy queries, etc. Retention times could provide valuable information in LC-MS based metabolomics but are comparable only over a certain range. Even when using the same column and mobile phase chemistries and column dimensions, retention times can vary between labs due to different LC instrumentation. The concept of retention time indexing (RTI), already widely used in GC-MS, can help simplify the process by converting retention times to the dimensionless retention index, which is only dependent on stationary phase chemistry and mobile phase composition. Here we present our first preliminary results from a ring-trial using a novel RTI system based on a homologous series of substances purposely designed for LC-MS and several metabolite standards measured in 5 different laboratories using different LC-MS systems. |
---|