Pubtator3 Ext Tools

Configuration File: pubtator3_ext_tools.json Tool Type: Local Tools Count: 2

This page contains all tools defined in the pubtator3_ext_tools.json configuration file.

Available Tools

PubTator3_GetEntityRelations (Type: BaseRESTTool)

Query the NCBI PubTator3 relation knowledge graph for a normalized biomedical entity and return R…

PubTator3_GetEntityRelations tool specification

Tool Information:

  • Name: PubTator3_GetEntityRelations

  • Type: BaseRESTTool

  • Description: Query the NCBI PubTator3 relation knowledge graph for a normalized biomedical entity and return RANKED relation edges aggregated across all of PubMed. Given a single entity (e1), returns every related entity with the relation type (associate, cause, treat, prevent, inhibit, stimulate, interact, compare, drug_interact, negative_correlate, positive_correlate, etc.), the source/target entities, and the aggregate publication count supporting each edge. Edges are returned ranked by supporting publication count (highest first). This answers ‘what is entity X related to, ranked by literature evidence’ as a graph query, distinct from returning the articles themselves. Optionally pass a relation type and/or a second entity (e2) to query a specific pair (e.g. does chemical A treat disease B). Entities must use PubTator’s normalized @TYPE_Name format, e.g. @GENE_BRAF, @DISEASE_Melanoma, @CHEMICAL_Doxorubicin. Use PubTator3_EntityAutocomplete to resolve free text into these normalized identifiers.

Parameters:

  • e1 (string) (required) Primary normalized PubTator entity in @TYPE_Name format. Examples: ‘@GENE_BRAF’, ‘@DISEASE_Melanoma’, ‘@CHEMICAL_Doxorubicin’, ‘@CHEMICAL_Vemurafenib’. Required.

  • type ([‘string’, ‘null’]) (optional) Optional relation type filter to restrict edges. Examples: ‘treat’, ‘cause’, ‘associate’, ‘prevent’, ‘inhibit’, ‘stimulate’, ‘interact’, ‘compare’, ‘drug_interact’, ‘negative_correlate’, ‘positive_correlate’. Omit to return all relation types.

  • e2 ([‘string’, ‘null’]) (optional) Optional second normalized PubTator entity (@TYPE_Name) to query a specific pair relationship between e1 and e2, e.g. e1=’@CHEMICAL_Doxorubicin’, type=’treat’, e2=’@DISEASE_Neoplasms’. Omit for a one-entity neighborhood query.

  • limit ([‘integer’, ‘null’]) (optional) Maximum number of top-ranked relation edges to return (client-side truncation; the API may return thousands of edges). Default 50.

Example Usage:

query = {
    "name": "PubTator3_GetEntityRelations",
    "arguments": {
        "e1": "example_value"
    }
}
result = tu.run(query)

PubTator3_get_annotations (Type: BaseRESTTool)

Extract biomedical entity annotations from PubMed articles using NCBI PubTator3. Given one or mor…

PubTator3_get_annotations tool specification

Tool Information:

  • Name: PubTator3_get_annotations

  • Type: BaseRESTTool

  • Description: Extract biomedical entity annotations from PubMed articles using NCBI PubTator3. Given one or more PubMed IDs (PMIDs), returns text-mined entities including genes, diseases, chemicals, species, mutations, and cell lines with their normalized identifiers (NCBI Gene ID, MeSH, NCBI Taxonomy, etc.) and text positions. Useful for extracting structured biomedical knowledge from literature. Provide PMIDs as a comma-separated string.

Parameters:

  • pmids (string) (required) Comma-separated PubMed IDs (e.g., ‘33205991’, ‘33205991,34234088’). Maximum ~100 PMIDs per request.

  • concepts (string) (optional) Comma-separated entity types to extract: gene, disease, chemical, species, mutation, cellline (default: all types)

Example Usage:

query = {
    "name": "PubTator3_get_annotations",
    "arguments": {
        "pmids": "example_value"
    }
}
result = tu.run(query)