NP-KG: Knowledge Graph Framework for Natural Product-Drug Interactions

NP-KG is a biomedical knowledge graph (KG) to identify and generate mechanistic hypotheses for pharmacokinetic natural product-drug interactions (NPDIs). NP-KG uses the PheKnowLator ecosystem to create an ontology-grounded KG. It then uses two relation extraction systems to extract triples from natural product-related biomedical literature to create a literature-based graph, and integrates the nodes and edges in the ontology-grounded KG. See GitHub repository for more details and instructions: https://github.com/sanyabt/np-kg. <strong>File Descriptions:</strong> <strong>I. NP-KG Merged TSV format</strong> <em>PheKnowLator_machine_read_merged_instance_based_OWLNETS_v1.0.1.tsv - </em>Merged KG with ontology-grounded KG and literature-based graph as TSV file <em>nodeLabels_v1.0.1.tsv</em> <strong>II. NP-KG Merged gpickle format</strong> <em>PheKnowLator_machine_read_merged_instance_based_OWLNETS_v1.0.1.gpickle - </em>Merged KG with ontology-grounded KG and literature-based graph as NetworkX multidigraph object. <em>nodeLabels_v1.0.1.pickle - </em>Node labels for all nodes in the KG. <strong>III. PheKnowLator instance-based KG (v3.0.0)</strong> <em>PheKnowLator_v3.0.0_full_instance_inverseRelations_OWLNETS_NetworkxMultiDiGraph.gpickle - </em>PheKnowLator KG with full instance build v3.0.0 as NetworkX multidigraph object. <strong>IV. Literature-based Graph in gpickle and ntriples format</strong> <em>machine_read_merged_with_closure_v1.0.1.pickle - </em>Pickle file of literature-based graph <em>machine_read_merged_with_closure_v1.0.1.nt - </em>ntriples file of literature-based graph.