Reproducibility of mass spectrometry based metabolomics data

Area of Science:

• Biomedical Science
• Analytical Chemistry
• Bioinformatics

Background:

• Reproducibility of measurements is crucial for reliable high-throughput metabolomics data analysis.
• Metabolites are defined as reproducible if consistent across replicate experiments.
• Mass spectrometry-based metabolomics (MS-Metabolomics) experiments require robust methods for assessing reproducibility.

Purpose of the Study:

• To introduce and evaluate Maximum Rank Reproducibility (MaRR) for assessing reproducibility in MS-Metabolomics data.
• To examine metabolite reproducibility across technical and biological samples in three MS-Metabolomics datasets.
• To provide a data-driven approach for evaluating the reliability of metabolomics measurements.

Main Methods:

• MaRR is a nonparametric approach utilizing a maximal rank statistic to identify shifts from reproducible to irreproducible signals.
• MaRR does not rely on parametric assumptions about metabolite data distributions or dependence structures.
• The method was applied to three MS-Metabolomics datasets from the Genetic Epidemiology of Chronic Obstructive Pulmonary Disease (COPD) study.

Main Results:

• MaRR effectively controlled the False Discovery Rate (FDR) under realistic MS-Metabolomics data conditions.
• The procedure demonstrated high power for detecting reproducible metabolites, with minimal bias in estimating reproducible signal proportions.
• Real data analysis revealed higher reproducibility for technical replicates than for biological replicates across all three datasets.

Conclusions:

• The MaRR procedure is adaptable to various experimental designs in metabolomics.
• The nonparametric approach offers consistent performance for assessing MS-Metabolomics reproducibility.
• The methods are implemented in the open-source R package 'marr', available via Bioconductor.