Computer Techniques for Mass Spectral Identification

Publisher Website
Google Scholar
Abstract
Techniques generally available for computer identification of unknown mass spectra include the following retrieval systems: PEAK, a user interactive system for reference file interrogation; KB, in which the two most abundant peaks every 14 mass units are used for matching, as proposed by Biemann and his associates at MIT; and PBM, a Probability Based Matching system employing weighting of mass and abundance data and reverse searching. Tests with a large number of unknown mass spectra have demonstrated the superiority of PBM, especially for unknown spectra of mixtures. The only generally applicable computer system for interpretation of mass spectra is the Self-Training Interpretive and Retrieval System (STIRS). These systems are generally available to outside users over two international computer networking systems, MSDC/Cyphernet and Cornell/TYMNET. Most of the use of the former involves PEAK and KB, while the latter has only PBM and STIRS available. A variety of systems have been utilized for in-house use and in commercial GC/MS/computer systems. One of the latter, OLFAX, combines data acquisition with PBM matching to give real-time identification and quantitation for preselected compounds.

This publication has 0 references indexed in Scilit: