7.3.7 Library Searching

Once a usable spectrum is displayed, the database of library spectra can be searched. The search algorithm assigns a weight to each mass peak according to how common that mass occurs in thousands of mass spectra. For example, m/z 43 is quite common, so the presence of this signal in an unknown is less informative than say an m/z 100 (which is considerably less common). This type of weighting is called Probability Based Matching, or PBM. In addition, the search is a reverse search, which means the library spectral peaks are compared to the unknown rather than vice versa.

Library search results have questionable meaning due to:

1. quality of the spectra entered into the library
2. the nature of the search algorithm
3. fundamental instrument differences (commonly, sector vs. quadrupole analyzers)
4. tuning variations

Searching is used only as a tool to aid identification, never as an identification itself. A good library search result does not eliminate the need for proper spectra interpretation by the analyst.

Exercise: Select the apex spectrum of the peak at 6.4 minutes in the data file 00376.D Do a library search (by double clicking the right mouse button while the curson is anywhere within the spectrum window). Print the results of this search. Next, select an average spectrum for the same peak and repeat the search. Is there a difference in the search results? Why?

Exercise: From the data file 00677.D, select an average spectrum for the peak at 5.7 minutes and search the library for this spectrum. Does the 'best hit' identify the unknown compound? Which one of the hits is the unknown? Why did the library search fail in this instance? What does this exercise demonstrate?