{
  "schema_version": 1,
  "series_id": "clingen-gene-disease-validity",
  "retrieved_at": "2026-09-22",
  "data_through": "2026-09-17",
  "source": {
    "schema_version": 1,
    "retrieved_at": "2026-09-22",
    "sources": {
      "history_archive": {
        "url": "https://storage.googleapis.com/genegraph-stage-public/clingen-gene-validity-json-all.tar.gz",
        "bytes": 32212025,
        "sha256": "c0a87f3612e54ca1a3274a0ab4e518f9ad271dccfbeaaf596df4efbe7ee30998"
      },
      "latest_archive": {
        "url": "https://storage.googleapis.com/genegraph-stage-public/clingen-gene-validity-json-latest.tar.gz",
        "bytes": 21102303,
        "sha256": "e3d4b8b867ef64851bd19633c83a97779f5ca2169b8c9449d657a26229f08326"
      },
      "current_csv": {
        "url": "https://search.clinicalgenome.org/kb/gene-validity/download",
        "bytes": 1122207,
        "sha256": "8f30ae12bfb6843f1273ef3ce06e28ce98a0f953ed99369ea7138da586c5b5ea"
      }
    },
    "citation": "Wright MW, Thaxton CL, Nelson T, et al. Generating Clinical-Grade Gene-Disease Validity Classifications Through the ClinGen Data Platforms. Annu Rev Biomed Data Sci. 2024;7:31-50.",
    "reuse_note": "The audit uses official public ClinGen and GeneGraph downloads and publishes derived tables plus source hashes, not the bulk archives."
  },
  "history": {
    "all_archive": {
      "members": 7624,
      "substantive": 6425,
      "non_substantive": 1199
    },
    "latest_archive": {
      "members": 3614,
      "substantive": 3614
    },
    "substantive_versions": 6425,
    "latest_base_assertions": 3614,
    "relationships": 4061,
    "classification_counts": {
      "Limited": 924,
      "Moderate": 903,
      "Definitive": 3811,
      "No Known Disease Relationship": 166,
      "Disputed": 345,
      "Strong": 214,
      "Refuted": 62
    },
    "current_classification_counts": {
      "Definitive": 2279,
      "Disputed": 188,
      "Limited": 533,
      "Refuted": 48,
      "Moderate": 452,
      "Strong": 79,
      "No Known Disease Relationship": 35
    },
    "approved_date_coverage": 1.0,
    "relationship_identity_coverage": 1.0,
    "current_label_mapping_coverage": 0.900545
  },
  "current_csv": {
    "header_index": 4,
    "headers": [
      "GENE SYMBOL",
      "GENE ID (HGNC)",
      "DISEASE LABEL",
      "DISEASE ID (MONDO)",
      "MOI",
      "SOP",
      "CLASSIFICATION",
      "ONLINE REPORT",
      "CLASSIFICATION DATE",
      "GCEP"
    ],
    "rows": 3678,
    "mapped_relationships": 3677,
    "duplicate_relationships": 1,
    "unmapped_moi": {},
    "file_created_line": "\"FILE CREATED: 2026-09-22\",\"\",\"\",\"\",\"\",\"\",\"\",\"\",\"\",\"\""
  },
  "outcomes": {
    "first_high_confidence_relationships": 2624,
    "currently_high_confidence_relationships": 2358,
    "high_confidence_exits": 20,
    "first_high_confidence_by_month": {
      "2016-12": 1,
      "2017-02": 1,
      "2017-05": 1,
      "2017-06": 10,
      "2017-08": 2,
      "2017-09": 4,
      "2017-10": 10,
      "2017-11": 11,
      "2017-12": 12,
      "2018-01": 13,
      "2018-02": 10,
      "2018-03": 11,
      "2018-04": 15,
      "2018-05": 19,
      "2018-06": 22,
      "2018-07": 17,
      "2018-08": 8,
      "2018-09": 22,
      "2018-10": 17,
      "2018-11": 8,
      "2018-12": 14,
      "2019-01": 13,
      "2019-02": 10,
      "2019-03": 17,
      "2019-04": 25,
      "2019-05": 15,
      "2019-06": 18,
      "2019-07": 16,
      "2019-08": 14,
      "2019-09": 13,
      "2019-10": 14,
      "2019-11": 21,
      "2019-12": 10,
      "2020-01": 26,
      "2020-02": 21,
      "2020-03": 11,
      "2020-04": 20,
      "2020-05": 38,
      "2020-06": 15,
      "2020-07": 32,
      "2020-08": 25,
      "2020-09": 22,
      "2020-10": 28,
      "2020-11": 20,
      "2020-12": 22,
      "2021-01": 26,
      "2021-02": 23,
      "2021-03": 24,
      "2021-04": 22,
      "2021-05": 24,
      "2021-06": 37,
      "2021-07": 47,
      "2021-08": 16,
      "2021-09": 17,
      "2021-10": 28,
      "2021-11": 29,
      "2021-12": 23,
      "2022-01": 21,
      "2022-02": 34,
      "2022-03": 33,
      "2022-04": 30,
      "2022-05": 38,
      "2022-06": 48,
      "2022-07": 33,
      "2022-08": 28,
      "2022-09": 39,
      "2022-10": 26,
      "2022-11": 31,
      "2022-12": 36,
      "2023-01": 35,
      "2023-02": 26,
      "2023-03": 25,
      "2023-04": 36,
      "2023-05": 35,
      "2023-06": 42,
      "2023-07": 38,
      "2023-08": 30,
      "2023-09": 53,
      "2023-10": 29,
      "2023-11": 26,
      "2023-12": 41,
      "2024-01": 28,
      "2024-02": 30,
      "2024-03": 36,
      "2024-04": 29,
      "2024-05": 32,
      "2024-06": 40,
      "2024-07": 19,
      "2024-08": 42,
      "2024-09": 37,
      "2024-10": 38,
      "2024-11": 30,
      "2024-12": 27,
      "2025-01": 22,
      "2025-02": 29,
      "2025-03": 22,
      "2025-04": 31,
      "2025-05": 31,
      "2025-06": 27,
      "2025-07": 26,
      "2025-08": 16,
      "2025-09": 26,
      "2025-10": 21,
      "2025-11": 8,
      "2025-12": 18,
      "2026-01": 24,
      "2026-02": 18,
      "2026-03": 34,
      "2026-04": 14,
      "2026-05": 9,
      "2026-06": 10,
      "2026-07": 11,
      "2026-08": 15,
      "2026-09": 1
    },
    "observations": 40
  },
  "gates": {
    "full_version_history": {
      "status": "passed",
      "evidence": "6425 substantive version records parsed from the full historical archive."
    },
    "approval_dates": {
      "status": "passed",
      "evidence": "100.0% of substantive versions have an Approved contribution date."
    },
    "relationship_identity": {
      "status": "passed",
      "evidence": "100.0% of substantive versions contain gene, disease, and inheritance identifiers."
    },
    "current_state_archive": {
      "status": "passed",
      "evidence": "3614 latest base assertions parsed from the official latest archive."
    },
    "human_readable_labels": {
      "status": "passed",
      "evidence": "90.1% of historical versions map to the current CSV for labels and expert panels."
    },
    "rebuild_path": {
      "status": "passed",
      "evidence": "All source URLs, hashes, transformations, derived tables, observations, and checks are public."
    }
  },
  "definition": {
    "event": "First Approved Strong or Definitive classification for a unique gene–disease–mode-of-inheritance relationship.",
    "date": "Approved contribution date in the GeneGraph version record.",
    "denominator": "Start-of-quarter relationships previously classified below high confidence plus half of new relationship entries during the quarter.",
    "quality_dimensions": [
      "Strong versus Definitive",
      "expert panel",
      "initial curation versus revision",
      "later high-confidence exit"
    ]
  }
}