PrimeKG-Plus: a literature-derived expansion of a multimodal precision medicine knowledge graph

PrimeKG-Plus (build 20260529) is an updated and literature-augmented extension of the Precision Medicine Knowledge Graph (PrimeKG). It integrates data from updated biomedical databases and expert-validated relations extracted from the scientific literature for four rare neurological disorders: Canavan disease, Batten disease, Niemann–Pick disease type C, and Tay–Sachs disease. PrimeKG-Plus is designed to support AI-driven drug repurposing, drug–target interaction prediction, and biomedical link prediction, with a focus on rare disease applications. The release contains two 12-column edge-list graphs: • primekg_plus.csv — base graph from updated public biomedical databases only (no PubMed-derived edges); 129,317 nodes; 7,683,206 directed edges. • primekg_plus_rd.csv — literature-augmented build for 4 studied rare disorders mentioned above; all edges in primekg_plus.csv plus 550 novel expert-validated literature edges (7,683,756 directed edges total). The bundle also includes the published Original PrimeKG baseline (no_dup_kg.csv), supplementary tables S1–S9, literature curation inputs, rebuild scripts (MIT), and documentation. Literature-derived relations are released under CC-BY 4.0. Associated manuscript submitted to Scientific Data. Code: https://github.com/DSDD-UCPH/PrimeKG-Plus Original PrimeKG: https://doi.org/10.7910/DVN/IXA7BM