FORUM

FORUM constructs a Semantic Web Knowledge Graph (KG) that semantically links chemical entities and biomedical concepts to support interpretation of metabolomics signatures, biomarker discovery, and hypothesis generation.

Key Features:

• Knowledge Graph size: A KG comprising over 8 billion triples and more than 8 million relationships that encodes chemical–biomedical concept associations.
• Semantic linking of chemicals and concepts: Semantically connects chemical compounds and metabolites to biomedical concepts and pathological outcomes.
• Federated data integration: Aggregates statements from a federation of life science databases and literature repositories, including PubMed and ChEBI.
• Semantic Web framework: Built on Semantic Web technologies and linked-data principles for ontological representation.
• Ontological-based reasoning: Applies ontological reasoning to infer new relationships and provide multiple levels of abstraction.
• Enrichment analysis: Performs enrichment analysis to estimate the statistical relevance of both explicitly extracted and inferred relations.
• Triplestore and query access: Stores the KG in a triplestore and exposes a SPARQL endpoint for direct querying.
• FAIR compliance: Adheres to FAIR principles to promote findability, accessibility, interoperability, and reusability of linked data.

Scientific Applications:

• Metabolomics interpretation: Associates metabolic signatures (lists of metabolites) with biomedical concepts to interpret experimental metabolomics results.
• Biomarker discovery and classification: Supports identification of candidate biomarkers and classification of individuals based on metabolite signatures.
• Disease mechanism analysis: Links metabolic signatures to pathological outcomes to study biological processes and disease mechanisms.
• Hypothesis generation: Enables generation of novel hypotheses by integrating inferred relationships and multi-level abstractions.
• Literature-supported association extraction: Extracts and integrates literature evidence to support chemical–biomedical concept associations.

Methodology:

Construction of a Semantic Web Knowledge Graph from federated life science databases and PubMed/ChEBI sources, storage in a triplestore with a SPARQL endpoint, application of ontological-based reasoning to infer relations, and use of enrichment analysis to estimate statistical relevance of extracted and inferred relations.