MetabR: an R script for linear model analysis of quantitative metabolomic data

Area of Science:

• Systems Biology
• Metabolomics
• Bioinformatics

Background:

• Metabolomics, a high-throughput systems biology approach, lacks robust data analysis tools compared to transcriptomics and proteomics.
• Metabolomic data analysis requires normalization to account for confounding variables, which current tools may not handle effectively for all metabolite relationships or random effects.
• Linear mixed models offer a flexible and standardized method for addressing both fixed- and random-effect confounding variables in metabolomic data.

Purpose of the Study:

• To introduce MetabR, a user-friendly program for statistical analysis of metabolomic data.
• To implement linear mixed models for normalization of metabolomic data, addressing limitations of existing tools.
• To provide researchers without programming backgrounds with a tool for analyzing complex metabolomic datasets.

Main Methods:

• Development of MetabR, a menu-driven program in the R statistical language.
• Implementation of linear mixed models for normalization of metabolomic data.
• Utilization of analysis of variance (ANOVA) for testing treatment differences and identifying differentially abundant metabolites.

Main Results:

• MetabR successfully normalizes metabolomic data using linear mixed models, accounting for various confounding variables.
• The program facilitates the identification of differentially abundant metabolites and aids in data interpretation.
• Example data demonstrates the program's utility in handling common confounding variables in metabolomic studies.

Conclusions:

• MetabR provides a simple, user-friendly solution for normalization and statistical analysis of targeted metabolomic data.
• The tool helps bridge the gap in available data analysis software for the metabolomics field.
• The MetabR program, documentation, and example data are available online for researchers.