1C-O2-1800(1P-16) PDF
A strategy based on spectral network for database-assisted identification of metabolites
A Database-assisted spectrum matching approach is widely used to systematically annotate metabolites in various types of LC-MS/MS-based metabolomics studies. However, it is still difficult to achieve highly sensitive and reliable chemical identification using ordinary spectrum-matching approach. In order to achieve sensitive chemical identifications of metabolites, we have developed a new approach by using spectral network strategy. We focused on the fact that structurally related metabolites derived from same metabolic pathways are simultaneously observed in LC-MS/MS analysis of biological matrices. Unlike ordinary database-assisted spectrum matching, our approach builds “spectral network” which is a network consisting of fragmentation spectra based on the similarities of fragmentation patterns. Chemical identification is determined by scoring chemical similarities between candidate structures in the network. Our method is especially effective for LC-MS/MS data containing large number of structurally related metabolites, such as secondary metabolites in plants. Also this method can be used complementarily to ordinary spectrum matching approach.