Background Large-scale gas chromatography-mass spectrometry (GC-MS) -based untargeted metabolomics, where hundreds or thousands of samples are analysed over weeks or months, has specific challenges. In addition to the computational challenges of compound identification that are intensified when dealing with a large number of samples, several sources of data variation are present. These include variation in instrument performance, signal intensity loss due to column ageing, the build-up of contaminants in the ion source, and sample handling variability. . Therefore, a data processing software package to address these problems is required. Results Our software, MassOmics, is designed to bring together R packages and scripts for GC-MS data processing to rapidly integrate and annotate peaks in large-scale datasets, all within a graphical user interface. This package also provides identification of background contaminants, data scaling and transformation, various batch effect removal methods, machine learning-powered grouping of metabolites, and metabolite importance analysis. With these functions, MassOmics can parse and summarise library batch search results from ChemStation/ MassHunter and produce an integrated output of the GC-MS datasets, which is compatible with various downstream statistical and metabolic pathway analysis tools. The module-based design and intermediate data transferring approach enable MassOmics to work with data integration platforms such as KNIME to generate an adaptive and customizable processing workflow. Conclusion The MassOmics package is designed for researchers with little experience using R, and substantially improves GC-MS data extraction efficiency and accuracy, as well as reducing the time required for manual checking and re-integration.

Funding support: Mass Spectrometry Hub, a Strategic Research Initiative at the University of Auckland