IPO: a tool for automated optimization of XCMS parameters

Area of Science:

• Metabolomics
• Bioinformatics
• Analytical Chemistry

Background:

• Untargeted metabolomics generates vast datasets requiring automated processing.
• Existing software packages are sensitive to algorithm parameter settings, risking biased results.
• A need exists for a fast, widely applicable parameter optimization tool free from complex labeling.

Purpose of the Study:

• To develop and validate a software package for optimizing metabolomics data processing parameters.
• To improve the accuracy and reliability of untargeted metabolomics data analysis.

Main Methods:

• Implemented Isotopologue Parameter Optimization (IPO) software using stable (13)C isotopic peaks for peak picking.
• Optimized retention time correction by minimizing relative differences within peak groups.
• Utilized design of experiments and response surface models for parameter optimization.

Main Results:

• IPO increased reliable metabolite groups by 146%-361% and decreased non-reliable groups to 3%-8%.
• Achieved a one-third reduction in retention time deviation.
• Demonstrated applicability across diverse sample types and liquid chromatography-high resolution mass spectrometry methods.

Conclusions:

• IPO successfully enhances the reliability of metabolomics data from liquid chromatography-high resolution mass spectrometry.
• The software is freely available, implemented in R, and tested on Linux/Windows.
• IPO offers a valuable tool for optimizing untargeted metabolomics data processing.