You are here:
The LCMS-Based Untargeted Metabolomics Assistant (LUMA) R package
Citation:
Evich, M., W. Melendez, AND J. Mosley. The LCMS-Based Untargeted Metabolomics Assistant (LUMA) R package. U.S. Environmental Protection Agency, Washington, DC, 2020.
Impact/Purpose:
Advances in high-resolution liquid chromatography mass spectrometry (HR LC-MS) have allowed for the detection of hundreds to thousands of metabolite features in complex matrices, introducing data processing hurdles in the field of metabolomics. The generation of such large datasets requires a need to simplify biologically relevant interpretation and introduce a degree of automation, to maximize reproducibility and minimize user input. Current metabolomics data processing tools, including XCMS and CAMERA exist to aid in data processing; while these tools are widely accepted, the need to learn a programing language (R) remains a hurdle for many novice metabolomics scientists. We introduce an R-based package, LUMA (LCMS-Based Untargeted Metabolomics Assistant) to bridge the current gaps between data collection and biological interpretation. The LUMA tool introduces modules, which contain functions that utilize existing data processing packages (XCMS and CAMERA) without an in-depth prerequisite knowledge of R programing. LUMA expedites the annotation and data curation step of CAMERA by consolidating visualization outputs of CAMERA for all features attributed to a single metabolite. These additions shorten the manual time spent visually inspecting data for potentially conflicting annotations. Lastly, LUMA can be used to simplify data by removing erroneous metabolite features to streamline biological interpretation. Together, this R package can be used by both novice and expert metabolomics scientists to streamline LC-MS data processing.
Description:
The LCMS-Based Untargeted Metabolomics Assistant (LUMA) package is an automated data processing tool, bridging the outputs of XCMS and CAMERA to a comprehensive data matrix in the R environment for discovery-based studies. The LUMA package performs feature-reduction by minimizing single features which do not pass quality control and negatively impact downstream statistical analyses. LUMA contains functions that allow for rapid, automated workflows to perform these steps with minimal user input. Furthermore, to expedite manual data curation for potentially conflicting isotope and ion adduct annotations, data visualization is consolidated to a single graphic per metabolite group. This graphic contains all EIC plots and psSpectra from CAMERA and new correlation matrices and dendrograms for all features attributed to a single metabolite. Final processed metabolite data, containing normalized intensities and user-defined metadata, can be exported to worksheets which are directly formatted to a number of analytical tools including Simca and MetaboAnalyst/Mummichog, while retaining traceability in the R environment.
