Mapping out the Metabolome: Metabolite Annotation Propagation and Synthesis (MAPS)

Identification and annotation of unknown metabolites remain a significant challenge in untargeted metabolomics due to the vast chemical diversity and limitations of individual annotation tools. The putative identification of metabolites frequently yields multiple candidate matches, and systematically reviewing annotations from databases such as GNPS and HMDB to determine the most accurate assignment can be a labor-intensive process. Here, we present MAPS (Metabolite Annotation Propagation and Synthesis), an automated annotation workflow that integrates, refines, and extrapolates metabolite annotations from multiple sources within a user-friendly R Shiny application. MAPS leverages the strengths of various data processing and annotation tools used in metabolomics, including MZmine, GNPS2, SIRIUS, and MS2Query, and employs an annotation propagation strategy to enhance coverage, as well as synthesizing annotations across spectral matching and cheminformatic tools to increase annotation confidence, and provide the most exhaustive list of high-confidence metabolites present. One significant advantage to MAPS is its ability to perform compound name standardization, which is not present in any other tool of its kind. To demonstrate its utility, we apply MAPS towards the curation of metabolites from the human gut metabolome; a complex metabolic system for which knowledge-based metabolome annotations tools are currently incapable of interrogating. Such a tool empowers deep metabolome mining in a more accessible, reliable and reproducible manner.