OptimusKG: A multimodal biomedical knowledge graph from structured resources

OptimusKG has been launched by researchers as a multimodal biomedical labeled property graph (LPG) that merges both structured and semi-structured data sources. This integration ensures the retention of factual, type-specific metadata across various domains, including molecular, anatomical, clinical, and environmental fields. The graph comprises 190,531 nodes categorized into 10 entity types, 21,813,816 edges representing 26 relation types, and 67,249,863 property instances that encode 110,276,843 values across 150 unique property keys, sourced from 18 ontologies and controlled vocabularies. OptimusKG implements a top-level schema for its nodes and edges while preserving detailed, type-specific properties, cross-references, and provenance, effectively addressing the challenge of unifying diverse biomedical resources.