RaMP: A Comprehensive Relational Database of Metabolomics Pathways for Pathway Enrichment Analysis of Genes and Metabolites

Abstract

The value of metabolomics in translational research is undeniable, and metabolomics data are increasingly generated in large cohorts. The functional interpretation of disease-associated metabolites though is difficult, and the biological mechanisms that underlie cell type or disease-specific metabolomics profiles are oftentimes unknown. To help fully exploit metabolomics data and to aid in its interpretation, analysis of metabolomics data with other complementary omics data, including transcriptomics, is helpful. To facilitate such analyses at a pathway level, we have developed RaMP (Relational database of Metabolomics Pathways), which combines biological pathways from the Kyoto Encyclopedia of Genes and Genomes (KEGG), Reactome, WikiPathways, and the Human Metabolome DataBase (HMDB). To the best of our knowledge, an off-the-shelf, public database that maps genes and metabolites to biochemical/disease pathways and can readily be integrated into other existing software is currently lacking. For consistent and comprehensive analysis, RaMP enables batch and complex queries (e.g., list all metabolites involved in glycolysis and lung cancer), can readily be integrated into pathway analysis tools, and supports pathway overrepresentation analysis given a list of genes and/or metabolites of interest. For usability, we have developed a RaMP R package (), including a user-friendly RShiny web application, that supports basic simple and batch queries, pathway overrepresentation analysis given a list of genes or metabolites of interest, and network visualization of gene-metabolite relationships. The package also includes the raw database file (mysql dump), thereby providing a stand-alone downloadable framework for public use and integration with other tools. In addition, the Python code needed to recreate the database on another system is also publicly available (). Updates for databases in RaMP will be checked multiple times a year and RaMP will be updated accordingly.

MDM2-Dependent Rewiring of Metabolomic and Lipidomic Profiles in Dedifferentiated Liposarcoma Models

“…The six metabolites altered by MDM2 binding inhibition in MDM2 lower cells comprised three lipids (1-(1-enyl-palmitoyl)-2-oleoyl-Glycerophosphoethanolamine (P-16:0/18:1), myristate (14:0), palmitoylcholine) and three nucleotides (adenylosuccinate, thymidine 5 -monophosphate, uridine). Pathways associated with these metabolites in the RaMPpathway database [ 12 ] included pyrimidine salvage, pyrimidine catabolism, and nucleotide salvage.…”

“…We developed a novel network-based pathway enrichment strategy that integrates pathway annotations from the RaMP database [12], as well as chemical structure similarity. Metabolites associated with phenotype (e.g., MDM2 amplification or treatment) that mapped to KEGG pathways [48][49][50] were used as the input.…”

“…(d) Over-represented pathways (FDR-adjusted p-value < 0.05) in metabolites that are altered between cells with higher vs. lower MDM2 amplification in the blue module of the final network enrichment model (Figure S1c). Pathways are grouped by cluster as determined by the relational database of metabolomics pathways (RaMP) pathway clustering algorithm[12].…”

Tools and resources for metabolomics research community: A 2017–2018 update

“…Relational database of Metabolomics Pathways (RaMP) , implemented as an R package, combines biological pathways from KEGG, Reactome, WikiPathways, and HMDB to aid in integration of metabolomics data with other omics datasets . Further, it enables batch and complex queries and can readily be integrated into pathway analysis tools and supports pathway overrepresentation analysis given a list of genes and/or metabolites of interest.…”

PathMe: Merging and exploring mechanistic pathway knowledge

“…OmniPath ( Türei et al ., 2016) have been created with the primary intention to integrate pathway knowledge from multiple databases. Beyond these, other approaches such as those taken in PathCards (Belenky et al ., 2015), RaMP (Zhang et al ., 2018), and ComPath (Domingo-Fernández et al, 2019), have focused on integrating gene sets and chemical knowledge related to pathways, but without including their topological information (i.e, relationships were excluded from the network). For instance, ComPath, the precursor of this work, harmonized pathway information at the gene level in order to conduct extensive manual curation that mapped and cross-referenced pathway representations across databases.…”