Comparative evaluations of mass spectral data bases.

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Data Science

Background:

• The National Institute of Science and Technology (NIST) maintains a large collection of mass spectra.
• NIST reports its collection is unique due to the completeness of its spectra.
• Mass spectral data is crucial for compound identification and analysis.

Purpose of the Study:

• To compare the comprehensiveness and performance of the 1989 Registry of Mass Spectral Data against the NIST database.
• To evaluate the effectiveness of spectral matching algorithms using both datasets.

Main Methods:

• Analysis of spectral data from the 1989 Registry (139,859 spectra) and the NIST database (53,994 spectra).
• Comparison of the average number of peaks per spectrum.
• Evaluation of spectral matching accuracy using a 68% reliability criterion for compounds present in both databases.

Main Results:

• The 1989 Registry's subset of spectra with the most peaks averaged 108 peaks per spectrum, 48% more than the NIST database.
• Spectral matching recall was significantly higher with the Registry (36%) compared to NIST (14%) for common compounds.

Conclusions:

• The 1989 Registry of Mass Spectral Data provides a more detailed and effective resource for mass spectral analysis than the NIST database.
• The enhanced spectral information in the Registry leads to improved accuracy in identifying unknown compounds.