pubtator3_relations — PubTator3 biomedical relations mined from PubMed
- Description
A directed network of biomedical concepts (genes, chemicals, diseases, genetic variants, species) and their relations, mined from the scientific literature by PubTator3 of the NCBI, which annotates all PubMed abstracts (about 36 million) and the full texts of the PMC Open Access Subset with deep-learning text mining tools: AIONER for the recognition of the entities, GNorm2, TaggerOne and NLM-Chem for their normalization to database identifiers, tmVar3 for the variants, and BioREx for the relations between them. Each edge is a relation found in an article, from the first to the second concept of the relation, with the type of relation and the PubMed identifier of the article; two concepts are joined by one edge per article (and type) in which their relation was found, so that the network is a multigraph.
The types of relation are
associate,cause,compare,cotreat,drug_interact,inhibit,interact,negative_correlate,positive_correlate,prevent,stimulateandtreat. The concepts are identified by the databases used by PubTator3: NCBI Gene identifiers for genes, MeSH identifiers for chemicals and diseases, dbSNP identifiers or tmVar normalized names for variants, NCBI Taxonomy identifiers for species. The names of the concepts are not included (they are in the much larger annotation files of PubTator3).PubTator3 updates its files monthly, replacing them in place; the data are those of the version of the download date (the release of October 6, 2026, at the time of writing).
The network has the following properties:
- (Node)
concept: the concept, with its type and identifier, as given by PubTator3 (e.g.Chemical|MESH:D003911,Gene|7157). - (Node)
kind: the type of the concept:Gene,Chemical,Disease,Species,DNAMutation,ProteinMutation,SNP, ... - (Edge)
type: the type of the relation. - (Edge)
pmid: the PubMed identifier of the article.
- (Node)
- Tags
- Biological Informational Unweighted Multigraph Metadata
- Added
- 2026-10-31
- Citation
-
- C.-H. Wei, A. Allot, P.-T. Lai, R. Leaman, S. Tian, L. Luo, Q. Jin, Z. Wang, Q. Chen, Z. Lu, "PubTator 3.0: an AI-powered literature resource for unlocking biomedical knowledge." Nucleic Acids Research 52, W540-W546 (2024) 10.1093/nar/gkae235 Sci-Hub
- Upstream URL OK
-
https://www.ncbi.nlm.nih.gov/research/pubtator3/
https://ftp.ncbi.nlm.nih.gov/pub/lu/PubTator3/ - Upstream license
- Public domain (United States Government Work), https://ftp.ncbi.nlm.nih.gov/pub/lu/PubTator3/README.txt
- Networks
-
Tip: hover over a column header for a description.
Column legend
- $\left<k\right>$
- Mean degree
- $\sigma_k$
- Standard deviation of the degrees
- $\lambda_h$
- Largest eigenvalue of the Hashimoto (non-backtracking) matrix
- $\tau$
- Random walk mixing time
- $r$
- Degree assortativity
- $c$
- Global clustering coefficient
- $\oslash$
- (Pseudo-) diameter
- $S$
- Relative size of the largest component
- NPs
- Node properties
- EPs
- Edge properties