ROIMCR: a powerful analysis strategy for LC-MS metabolomic datasets

Area of Science:

• Chemometrics
• Metabolomics
• Mass Spectrometry

Background:

• Liquid chromatography-mass spectrometry (LC-MS) metabolomic data analysis is complex due to large data volumes.
• Existing methods often require chromatographic alignment and peak shaping, which can be limiting.
• A need exists for alternative LC-MS data analysis strategies applicable to diverse MS datasets.

Purpose of the Study:

• To present ROIMCR, an alternative approach for LC-MS data analysis.
• To demonstrate ROIMCR's ability to filter, compress, and resolve LC-MS datasets.
• To provide a detailed description of the ROIMCR method and its implementation.

Main Methods:

• ROIMCR (Regions of Interest Multivariate Curve Resolution-Alternating Least Squares) method development.
• Data filtering and compression via searching regions of interest (ROIs) in the m/z domain.
• Multivariate Curve Resolution-Alternating Least Squares (MCR-ALS) analysis for resolving compressed data without alignment or peak shaping.

Main Results:

• ROIMCR effectively filters and compresses massive LC-MS datasets into a feature matrix.
• The method resolves compressed data to identify contributing pure components.
• Analysis was performed using MATLAB, with examples from lipidomic and other LC-MS studies.

Conclusions:

• ROIMCR offers data filtering and compression with preserved spectral accuracy.
• It leverages MCR-ALS for data resolution without the need for chromatographic alignment or modeling.
• ROIMCR is a powerful alternative for untargeted metabolomics LC-MS data analysis, outperforming other MCR-ALS applications.