Compound identification using partial and semipartial correlations for gas chromatography-mass spectrometry data

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Data Analysis

Background:

• Compound identification is crucial for gas chromatography-mass spectrometry (GC-MS) data analysis.
• Mass spectrum matching using dot product is the current standard for identification.
• Transformations of fragment ion intensities are used to enhance identification accuracy.

Purpose of the Study:

• To introduce partial and semipartial correlations as novel spectral similarity measures for GC-MS compound identification.
• To evaluate the effectiveness of these measures with various peak intensity transformations.
• To develop mixture versions of the proposed methods to further improve identification accuracy.

Main Methods:

• Partial and semipartial correlations were employed as spectral similarity measures.
• Various transformations of peak intensity were applied.
• Mixture versions of the correlation methods were developed.
• The National Institute of Standards and Technology (NIST) mass spectral library was used for validation.

Main Results:

• The proposed mixture partial and semipartial correlations outperformed the standard dot product and its composite measure.
• The mixture similarity using semipartial correlation achieved the highest accuracy of 84.6%.
• Optimal transformation parameters for fragment ion intensity and m/z value were identified as (0.53, 1.3).

Conclusions:

• Partial and semipartial correlations represent a significant advancement in GC-MS compound identification.
• The developed mixture similarity measures offer superior accuracy compared to existing methods.
• This study provides a more robust approach for analyzing mass spectral data.