MEtabolism Semantic Atlas MeSA

MeSA (MEtabolism Semantic Atlas) is an interlinked atlas of RDF quotient graphs that summarizes the structural organization of metabolism-related semantic resources at large scale.

MeSA provides a compact representation of heterogeneous RDF knowledge graphs by grouping resources according to the properties they share, using a transitive property-co-occurrence approach derived from RDFQuotient. Each source graph is summarized independently while preserving its structural patterns, provenance, and representative information.

The resulting summaries are then connected across datasets: quotient classes from different graphs are linked when they contain shared IRIs, making explicit the structural and semantic connections between otherwise independent resources. These cross-graph correspondences are represented using skos:closeMatch.

Together, the graphs currently represented in MeSA account for more than 29 billion RDF triples. The atlas therefore provides a lightweight view over a data landscape that would otherwise be impractical to explore globally. It can be used to inspect the organization and overlap of metabolism-related semantic resources, identify relevant datasets and structural paths, and support query planning across them.

When the underlying datasets are exposed through SPARQL endpoints, MeSA can be used to guide federated SPARQL queries across the corresponding resources. The same approach can also be applied directly to RDF dumps, allowing large-scale querying and exploration in data-lake environments without requiring every source to provide a public endpoint.

MeSA is distributed itself as RDF, together with an interactive visualization for exploring source graphs, quotient classes, structural relations, and cross-dataset links.