Comprehensive Peak Characterization (CPC) in Untargeted LC-MS Analysis.

Area of Science:

• Analytical Chemistry
• Biochemistry
• Bioinformatics

Background:

• Untargeted metabolomics using LC-MS generates large, complex datasets.
• Automated peak detection algorithms often include false positives to avoid missing real signals.
• Filtering these artifacts is crucial for accurate data analysis.

Purpose of the Study:

• To present the CPC algorithm for automated peak characterization and quality filtering.
• To reduce false positive peaks in LC-MS metabolomics data.
• To enhance the reliability of downstream data processing and analysis.

Main Methods:

• Development of the CPC (Characterization of Peak Components) algorithm.
• Automated peak characterization using analytical chemistry quality criteria.
• Application and integration with XCMS workflow for LC-MS data.

Main Results:

• The CPC algorithm effectively characterizes detected peaks.
• Approximately 35% of XCMS-detected peaks were removed, primarily those with low signal-to-noise ratios.
• The algorithm successfully filtered low-quality peaks, reducing data complexity.

Conclusions:

• The CPC algorithm provides automated, reliable filtering of artifact peaks in metabolomics.
• It significantly improves data quality and reduces manual curation efforts.
• The R-package facilitates seamless integration into existing metabolomics workflows.