The Biomarker Knowledge Graph (Biomarker KG) is a biomedical RDF knowledge graph for organizing biomarker-centered evidence and analysis results. Its schema, reflected in the attached entity file, is grounded in Biolink and VoID terms, with core classes for datasets, studies, diseases, phenotypic features, disease–phenotype associations, entity–disease associations, and statistical analysis outputs such as concept-count and log-odds analysis results. The graph captures not only biomarker-related entities and associations, but also provenance and evidence: studies connect to analysis results, associations link to supporting studies, and results reference supporting datasets, sample sizes, counts, odds ratios, and confidence intervals. Public KG registry information describes BiomarkerKG as a resource integrating biomarker data across anatomical structures, compounds, conditions, roles, and variants to support exploration of biomarker relationships. Together, these design choices make Biomarker KG useful for biomarker discovery, cohort-level evidence review, disease-phenotype interpretation, and downstream graph analytics or machine-learning workflows.
