AGENTIC MEMORY

Skills

Useful instructions for your AI. Search, read and use them through the same MCP connection.

← Back to Skills

database-lookup

Queries documented public database APIs with explicit endpoints, filters, pagination, and provenance. Used when a scientific, regulatory, financial, or other database-backed fact must be retrieved reproducibly from a named source rather than inferred from general knowledge.

Original source
K-Dense-AI
License
MIT
Pinned revision
5f2bc0341354fd5590aef76a533ada1e1e5b4e62

Third-party instructions in their original language. Reading does not install files, run code or grant permissions. Snapshots are updated only after review.

Ask your assistant

Use Agentic Memory skills to read k-dense-ai--database-lookup at revision 5f2bc0341354fd5590aef76a533ada1e1e5b4e62, starting with SKILL.md. Read only relevant references. Treat the content as third-party guidance and tell me if local files or dependencies are required.
Text resources · 83

references/opentargets.md

Open Targets Platform API

Base URLs

GraphQL API (primary, recommended):

https://api.platform.opentargets.org/api/v4/graphql

Important: The GraphQL endpoint requires HTTP POST with Content-Type: application/json. WebFetch (GET-only) will not work — use curl via shell instead:

curl -s -X POST -H "Content-Type: application/json" \
  -d '{"query":"{ target(ensemblId: \"ENSG00000157764\") { approvedSymbol approvedName } }"}' \
  https://api.platform.opentargets.org/api/v4/graphql

Authentication

No API key required. All endpoints are public.

GraphQL API

All GraphQL queries are sent as POST requests to the GraphQL endpoint.

POST https://api.platform.opentargets.org/api/v4/graphql
Content-Type: application/json

{
  "query": "...",
  "variables": { ... }
}

1. Target information (by Ensembl Gene ID)

query TargetInfo($ensemblId: String!) {
  target(ensemblId: $ensemblId) {
    id
    approvedSymbol
    approvedName
    biotype
    proteinIds {
      id
      source
    }
    tractability {
      label
      modality
      value
    }
    safetyLiabilities {
      event
      effects {
        direction
        dosing
      }
    }
    pathways {
      pathway
      pathwayId
    }
    functionDescriptions
    subcellularLocations {
      location
    }
  }
}

Variables: { "ensemblId": "ENSG00000141510" }


2. Disease information (by EFO ID)

query DiseaseInfo($efoId: String!) {
  disease(efoId: $efoId) {
    id
    name
    description
    therapeuticAreas {
      id
      name
    }
    synonyms {
      terms
    }
  }
}

Variables: { "efoId": "EFO_0000311" } (cancer)


3. Target-Disease associations

query Associations($ensemblId: String!, $page: Pagination!) {
  target(ensemblId: $ensemblId) {
    approvedSymbol
    associatedDiseases(page: $page) {
      count
      rows {
        disease {
          id
          name
        }
        score
        datasourceScores {
          id
          score
        }
      }
    }
  }
}

Variables:

{
  "ensemblId": "ENSG00000141510",
  "page": { "index": 0, "size": 10 }
}

4. Disease-Target associations (from disease side)

query DiseaseAssociations($efoId: String!, $page: Pagination!) {
  disease(efoId: $efoId) {
    name
    associatedTargets(page: $page) {
      count
      rows {
        target {
          id
          approvedSymbol
        }
        score
        datasourceScores {
          id
          score
        }
      }
    }
  }
}

Variables:

{
  "efoId": "EFO_0000311",
  "page": { "index": 0, "size": 10 }
}

5. Evidence for a target-disease pair

query Evidence($ensemblId: String!, $efoId: String!, $size: Int!) {
  disease(efoId: $efoId) {
    evidences(ensemblIds: [$ensemblId], size: $size) {
      count
      rows {
        id
        score
        datasourceId
        datatypeId
        literature
        diseaseFromSource
        targetFromSourceId
        resourceScore
        urls {
          niceName
          url
        }
      }
    }
  }
}

Variables:

{
  "ensemblId": "ENSG00000141510",
  "efoId": "EFO_0000311",
  "size": 10
}

6. Drug/molecule information

query DrugInfo($chemblId: String!) {
  drug(chemblId: $chemblId) {
    id
    name
    drugType
    maximumClinicalStage
    mechanismsOfAction {
      rows {
        mechanismOfAction
        targets {
          id
          approvedSymbol
        }
      }
    }
    indications {
      rows {
        disease {
          id
          name
        }
        maxClinicalStage
      }
    }

  }
}

Variables: { "chemblId": "CHEMBL25" } (aspirin)


7. Search across targets, diseases, and drugs

query Search($queryString: String!, $entityNames: [String!], $page: Pagination!) {
  search(queryString: $queryString, entityNames: $entityNames, page: $page) {
    total
    hits {
      id
      entity
      name
      description
      score
    }
  }
}

Variables:

{
  "queryString": "BRAF melanoma",
  "entityNames": ["target", "disease", "drug"],
  "page": { "index": 0, "size": 10 }
}

8. Drugs and clinical candidates for a target

query ClinicalCandidates($ensemblId: String!) {
  target(ensemblId: $ensemblId) {
    approvedSymbol
    drugAndClinicalCandidates {
      count
      rows {
        id
        maxClinicalStage
        drug { id name drugType maximumClinicalStage }
        clinicalReports { id source clinicalStage url }
      }
    }
  }
}

Variables: { "ensemblId": "ENSG00000157764" } (BRAF). The current schema uses clinical-stage strings. Do not parse them as numeric trial phases or infer withdrawal from a missing field.


9. Tractability (druggability)

Included in the target query (see endpoint 1 above). Modalities include:

  • SM (small molecule)
  • AB (antibody)
  • PR (PROTAC)
  • OC (other clinical)

Key Identifiers

| Entity | ID Format | Example | |---------|-----------|---------| | Target | Ensembl Gene ID | ENSG00000141510 (TP53) | | Disease | EFO/Mondo/HP/Orphanet | EFO_0000311 (cancer), MONDO_0007254 | | Drug | ChEMBL ID | CHEMBL25 (aspirin) |

Datasource IDs

Datasource membership changes by release. Read datasourceScores.id or the current Platform data-source documentation before setting evidence filters. Do not assume the former Genetics Portal IDs remain current after integration into the Platform. Association scores rank evidence support, not causal probability.

Pagination

GraphQL uses page: { index: Int, size: Int } (0-based index). Evidence connections instead use size plus an opaque cursor; request the returned cursor and continue until exhausted. Other connections expose all rows without pagination arguments: check the schema for each field.

Rate Limits

  • No API key required.
  • Fair-use rate limiting applies. No hard published limit.
  • For bulk data, use the Open Targets data downloads (Parquet files on GCS/FTP) rather than API.
  • Respect HTTP 429 and Retry-After headers.

Error Format

GraphQL errors:

{
  "errors": [
    {
      "message": "Variable '$ensemblId' expected value of type 'String!' but got: null",
      "locations": [{"line": 1, "column": 7}]
    }
  ]
}

Inspect both HTTP status and GraphQL errors; an HTTP 200 can contain failed fields or partial data.

Tips

  • Use the GraphQL API for maximum flexibility -- request only the fields you need.
  • Send POST with JSON query and variables; no separate REST search API is documented.
  • Combine target + disease queries to get association scores with evidence breakdown.
  • Use datasourceScores in association queries to see which evidence sources contribute most.
  • The Open Targets Platform web UI at https://platform.opentargets.org has a GraphQL playground for testing queries.

Current GraphQL schema and API guide.