Computer-assisted library search system for identification of unknown mass spectra

Abstract

The library search system described identifies a single component or two components in the unknown pure or mixture mass spectra by comparing them with a large data base of reference spectra. A preliminary search is based on the spectral interpretation and the main search is based on the probability of peak appearance. The performance of this system was tested on 254 pure spectra and 88 mixture spectra. The percentages of successful search by using the NIH/EPA/MSDC data base were 75% for the pure spectra, and 63% for both the first and second components in mixture spectra. The percentages of successful search improved to 94% for the first components of the mixture spectra and 77% for both first and second components of the mixture spectra, when conditions for measurement of reference spectra were the same as those for the unknown spectra.