Weighting Low-Intensity MS/MS Ions and m/z Frequency for Spectral Library Annotation

Area of Science:

• Metabolomics
• Mass Spectrometry
• Computational Chemistry

Background:

• Spectral similarity calculation is crucial for MS/MS data analysis in untargeted metabolomics.
• Accurate identification and annotation of compounds rely on effective spectral matching.
• Previous methods enhanced matching by increasing peak intensities at high m/z ranges.

Purpose of the Study:

• To evaluate the impact of weighting strategies on identifying structurally related compounds and performing spectral library searches.
• To propose a novel weighting approach considering m/z value frequency and low-peak intensity.
• To benchmark the proposed weighting method against existing approaches.

Main Methods:

• Applied weighting preprocessing to modified cosine, entropy, and fidelity distance metrics.
• Compared the proposed weighting approach with previously reported weights.
• Evaluated performance in identifying structurally similar compounds and annotating unknown spectra.

Main Results:

• Weighting-based preprocessing significantly aids in annotating unknown spectra and identifying structurally similar compounds.
• The proposed method, incorporating m/z frequency and low-peak intensity, demonstrates improved spectral matching.
• Identified specific spectral features where weighting might negatively impact performance.

Conclusions:

• Weighting preprocessing is a valuable technique for enhancing MS/MS spectral similarity calculations in metabolomics.
• The proposed frequency-aware and intensity-boosting weighting approach offers improved accuracy in compound identification.
• Understanding the limitations and specific spectral features where weights are detrimental is essential for optimal application.