Today, large mass spectral data bases containing several hundred thousand spectra are commercially available. For the identification of a mass spectrum from a sample analysis, e.g. a GC-MS run, most search algorithms rank the spectra found in a database according to a match quality factor based on visual match of measured and library spectra.
Whereas these algorithms work fine if there is at least one reference spectrum in the database for the sample component under examination, they often fail to deliver any useful information in the case where no reference spectrum is in the library, e. in the case of real unknowns.
A search algorithm that does not suffer from this limitation is the SISCOM similarity search used in the MassLib mass spectrometry software package.
The SISCOM search algorithm is based on the presence of characteristic masses and finds structurally related spectra even if visually similar matches are not available or if spectra are of limited quality which is typical in trace analysis.
By combining the SISCOM search for fragment masses with neutral loss and structural similarity searches, MassLib is an invaluable tool not only for identification of components with reference spectra, but also for structure elucidation of real unknowns.
MSP Kofel is based in Zollikofen, Switxerland. www.msp.ch
