Prepare HBP 3.0 NAR release metadata

.zenodo.json

@@ -0,0 +1,39 @@
+{
+  "title": "HeartBioPortal DataHub: HBP 3.0 NAR release",
+  "creators": [
+    {
+      "name": "HeartBioPortal contributors"
+    },
+    {
+      "name": "TBD: confirm author list before release"
+    }
+  ],
+  "description": "Data integration, provenance, artifact publishing, and serving-datamart tooling for the HeartBioPortal 3.0 NAR Database Issue manuscript release.",
+  "access_right": "open",
+  "upload_type": "software",
+  "license": "cc-by-4.0",
+  "keywords": [
+    "HeartBioPortal",
+    "cardiovascular genomics",
+    "data integration",
+    "provenance",
+    "NAR Database Issue"
+  ],
+  "related_identifiers": [
+    {
+      "identifier": "https://github.com/HeartBioPortal/DataHub",
+      "relation": "isSupplementTo",
+      "scheme": "url"
+    },
+    {
+      "identifier": "https://heartbioportal.org/",
+      "relation": "isSupplementTo",
+      "scheme": "url"
+    },
+    {
+      "identifier": "https://github.com/HeartBioPortal",
+      "relation": "isPartOf",
+      "scheme": "url"
+    }
+  ]
+}

ARTIFACT_MANIFEST.tsv

@@ -0,0 +1,11 @@
+artifact_name	hbp_layer	artifact_path_or_pattern	artifact_type	key_fields	record_count	gene_count	variant_count	source_count	build_version	schema_version	compressed	redistributable	license_notes	description
+population_frequency_datamart	Population-frequency context	datamart/population_frequency*; datamart/association_serving_slim.duckdb	tables/duckdb	rsid; allele; population_label; population_group; sample_size	TBD; verify from production QA	TBD	TBD; expected about 18.1M rsIDs if production QA confirms	37 if production QA confirms	v3.0.0-nar	TBD	mixed	requires review	requires source-license review	Approximate 594.3M source-specific observations was not locally verified; confirm before release.
+association_artifacts	Association and phenotype evidence	analyzed_data_unified/association/final/**; association/final/**	json/json.gz/duckdb	gene; variant_id; phenotype_path; p_value; ancestry; consequence; clinical_significance	TBD; verify from production QA	TBD	TBD	TBD	v3.0.0-nar	TBD	mixed	requires review	requires source-license review	Gene-level association artifacts, variant-index payloads, phenotype rollups, and serving summaries.
+structural_variant_gene_payloads	Structural-variant evidence	analyzed_data/dbvar/dbvar_structural_variants_nstd229.json.zip	json zip	gene; sv_id; sv_type; coordinates; clinical_significance	3072942	75192	3040582	1	v3.0.0-nar	structural_variant_legacy	yes	requires review	requires TOPMed/dbVar source-term review	Local dbVar nstd229 report verifies records, variants, and genes.
+exon_enriched_sv_artifacts	Structural-variant evidence	analyzed_data/dbvar/dbvar_structural_variants_nstd229.exons.json.zip	json zip	gene; sv_id; transcript_overlap; exon_overlap	TBD; verify from production QA	TBD	TBD	1	v3.0.0-nar	TBD	yes	requires review	requires TOPMed/dbVar source-term review	Exon-enriched structural-variant payloads derived from dbVar nstd229 and annotation overlap.
+protein_context_artifacts	Protein context	secondary_analyses/final/protein_context/*.json.gz	json.gz	gene; transcript_id; translation_id; protein_accession; feature_id	TBD; verify from production QA	5304 local untracked payload files observed; production expected value requires QA	TBD	Ensembl/UniProt/EBI/InterPro sources	v3.0.0-nar	TBD	yes	requires review	requires source attribution and terms review	Production target mentions about 66.9k isoforms and more than 3.2M feature annotations; not locally verified.
+gene_profile_artifacts	Gene profiles	secondary_analyses/final/gene_profile/*.json.gz; generated gene-profile paths	json/json.gz	gene; source_id; xref	TBD; verify from production QA	TBD	TBD	HGNC/NCBI/UniProt/GOA/Reactome/etc.	v3.0.0-nar	TBD	mixed	requires review	requires source-license review	Gene summary, nomenclature, ontology, pathway, and cross-reference payloads.
+expression_payloads	Expression layer	secondary_analyses/final/expression/*.json.gz; configured expression outputs	json/json.gz	gene; tissue; cell_type; expression_value	TBD; verify from production QA	TBD	TBD	TBD	v3.0.0-nar	TBD	mixed	requires review	requires source-license review	Expression payloads imported or transformed for HBP gene dossiers.
+shared_architecture_summaries	Shared-architecture layer	secondary_analyses/final/sga/*.json.gz; configured SGA outputs	json/json.gz	gene; cvd_phenotype; trait_phenotype; shared_variant_count	TBD; verify from production QA	TBD	TBD	inherits association source count	v3.0.0-nar	TBD	mixed	requires review	inherits association source restrictions	Derived overlap summaries between CVD and trait association variants.
+drug_discovery_payloads	Drug-discovery / Drugs & Compounds layer	drug-discovery output paths; per-gene drug payloads	json/json.gz	gene; molecule_id; molecule_name; target_id; source	TBD; reported 17128 records requires production QA	TBD; reported 1839 gene files requires production QA	TBD	Open Targets; DrugBank	v3.0.0-nar	TBD	mixed	requires review	DrugBank raw data license-restricted; Open Targets terms require confirmation	Reported 17,128 records, 1,839 gene files, and 1,454 unique molecules were not locally verified.
+guideline_derived_payloads	Clinical guidelines / guideline graph links	HCG/HCG-KG derived import paths; hcgkg_llm artifacts	json/graph/vector	gene; guideline_id; recommendation_id; evidence_class; evidence_level; snippet_id	TBD; verify from HCG/HCG-KG release QA	TBD	TBD	AHA/ACC guideline sources; future ESC	v3.0.0-nar	TBD	mixed	requires review	guideline source licenses require review	Gene-first guideline context consumed by HBP search_summary and guideline_detail when present.

BUILD_METADATA.json

@@ -0,0 +1,31 @@
+{
+  "project": "HeartBioPortal",
+  "component": "DataHub",
+  "release": "3.0.0-nar",
+  "created_at": "2026-05-18T00:00:00-04:00",
+  "git_commit": "784decace6458cd0791e8149e09eb76552a31bee",
+  "datahub_schema_version": "TBD; confirm before release",
+  "genome_assembly": "GRCh38 if applicable; confirm per artifact",
+  "major_layers": [
+    "association and phenotype evidence",
+    "population-frequency context",
+    "variant annotation",
+    "structural-variant evidence",
+    "protein context",
+    "gene profiles",
+    "clinical guideline graph links",
+    "drug-discovery / drugs and compounds",
+    "expression",
+    "shared genetic architecture"
+  ],
+  "source_manifest": "DATA_SOURCES.tsv",
+  "artifact_manifest": "ARTIFACT_MANIFEST.tsv",
+  "restricted_data_policy": "Do not redistribute controlled individual-level human data, credentials, restricted raw source data, raw DrugBank full database files, or license-uncertain third-party source files.",
+  "controlled_individual_level_data": "not redistributed",
+  "notes": [
+    "Update git_commit after final release commit is selected.",
+    "Counts marked TBD in ARTIFACT_MANIFEST.tsv require production QA confirmation.",
+    "Local dbVar nstd229 report verifies 3072942 records, 3040582 variants, and 75192 gene-level payloads.",
+    "DrugBank v5.1.12 raw data are license-restricted and should not be included unless redistribution permission is confirmed."
+  ]
+}

CHECKSUMS.txt

@@ -0,0 +1,18 @@
+ecff620da4e234325c5616ee5e8efd732e074748b23705c99a5cf6c06ac24451  .zenodo.json
+f2df420efc5b1962e21fc9f431714605873150300bdbf03b5f59ab300cff57b7  ARTIFACT_MANIFEST.tsv
+76544cb7d297551bfadcb8816d433bb8b8855ac917a2f144bd1e5a3c480a9da7  BUILD_METADATA.json
+a4435542f0fb7ca1b8dbd1c14861a2c4628b637483a31c72bb52de6bd9647bc8  CITATION.cff
+7509ecb44ad4137b71ec083a454f50cb4987bed4f04f3d20c3af2abb8139f83b  DATA_SOURCES.md
+04982f407a637edfc0736a1b74fc59bf991430872f858d9a6ef9e3ad62c76a9d  DATA_SOURCES.tsv
+fbf48fe05e2e55283171b4c7a90b6388cb3e82d44eb7eb5c70c75e0a2973714a  LICENSE
+4f23d36d7930cf3c6702de7135ecd7ef781235809b50c608f56c654c0ec8aba5  LICENSES.md
+8f2e1fa3f4d582335b1486d465da986836f26d45ea334333692e11bf1f8279ea  MANIFEST.md
+5b14a73d64211e75c7ea7cf988741a96727babc44e094041d261d91893dcfcd5  PROVENANCE_SCHEMA.md
+1dcdf166eb05e0a3c1214fd06904628ea4cb9b673304b929db1bf6103d5518c0  README.md
+ba24cd85ffe6811b29928c46268ee28106b340c828e1453530b87255224c8e86  RELEASE_NOTES.md
+96962af6c01ef76ffcc75697b328d0513e1ef75cc97a97b0ab5c4585e72248e6  docs/schemas/drug_discovery.md
+2984ea49a550a3d3b91e51a0346a5316ba1e3520eb884a55ed877925a09317c1  docs/schemas/gene_profile.md
+96b9ce0a18ab071c26c1270727b13a4cf45d8d4833842ea322c60e40cbf0b694  docs/schemas/guideline_signal.md
+1e90d2145fbc4049d2f8354f9c30eae04ccd2e41ab923348541ec1506aa032cf  docs/schemas/population_frequency.md
+febfe8f9b31139c063b8ac4a5953a1e646da363359697e707fc3c7122d9feba7  docs/schemas/protein_context.md
+7fb87293e8b1446941a9f5dcf3bc9d42c7d6fed242751fd7a86dd771efea7c73  docs/schemas/structural_variant.md

CITATION.cff

@@ -0,0 +1,18 @@
+cff-version: 1.2.0
+message: "If you use this software or data-processing workflow, please cite the HeartBioPortal 3.0 manuscript and this archived release."
+title: "HeartBioPortal DataHub"
+version: "3.0.0-nar"
+date-released: "2026-05-18"
+license: "CC-BY-4.0"
+repository-code: "https://github.com/HeartBioPortal/DataHub"
+url: "https://heartbioportal.org/"
+authors:
+  - name: "HeartBioPortal contributors"
+  - name: "TBD: confirm author list before release"
+abstract: "Data integration, provenance, artifact publishing, and serving-datamart tooling for the HeartBioPortal 3.0 NAR Database Issue manuscript release."
+keywords:
+  - HeartBioPortal
+  - cardiovascular genomics
+  - data integration
+  - provenance
+  - NAR Database Issue

DATA_SOURCES.md

@@ -0,0 +1,39 @@
+# Data Sources Summary
+
+`DATA_SOURCES.tsv` is the machine-readable source inventory for the HBP 3.0 NAR release. This companion file summarizes the source families by HBP layer.
+
+## Association and phenotype evidence
+
+DataHub normalizes association rows from HBP legacy CVD/trait layers, Million Veteran Program summary-statistics inputs when available, GWAS Catalog when included, and other association profiles into canonical gene, variant, phenotype, p-value, ancestry, consequence, clinical-significance, and provenance fields. The final artifacts are association JSON/JSON.GZ payloads, variant-index payloads, phenotype rollups, and serving DuckDB tables. Source provenance should preserve input file, source dataset, source version, phenotype path, rsID/variant ID, p-value, genome build, and transformation notes. Controlled or non-public inputs must not be redistributed.
+
+## Population-frequency context
+
+Population-frequency context is expected to combine source-specific allele-frequency observations from resources such as ALFA, gnomAD v4, 1000 Genomes, 1000 Genomes 30X, TOPMed-derived public frequency resources, PAGE, HGDP-CEPH, HapMap, ExAC, SGDP, and 38KJPN where present in the production build. DataHub harmonizes rsID, allele, population label, population group, sample size, study/resource, genome build, and source provenance into population-frequency datamarts. Production totals must be verified from QA before release; the approximate 594.3 million frequency observations across 18.1 million rsIDs and 37 resources are not hard-coded in `ARTIFACT_MANIFEST.tsv` because they were not locally verifiable.
+
+## Variant annotation
+
+Variant annotation uses dbSNP, ClinVar, Ensembl Variation, ClinGen where included, and source-specific legacy fields. DataHub preserves rsID/variant IDs as variant-level keys for chart aggregation and records clinical significance, variation type, most-severe-consequence-like fields, source record IDs, source dataset, genome build, and transformation steps. Source licenses follow the original providers.
+
+## Structural-variant evidence
+
+Structural-variant evidence is currently represented by dbVar nstd102/ClinVar structural-variant seed payloads and dbVar nstd229/TOPMed structural-variant call-set artifacts. DataHub normalizes source DB, study/submission, SV ID, SV type, coordinates, event length, clinical significance when present, gene overlap, transcript overlap, and exon overlap. The local nstd229 report verifies 3,072,942 records, 3,040,582 variants, and 75,192 gene-level payloads. TOPMed-related licensing and redistribution constraints require final review before public archival of source or derived bulk artifacts.
+
+## Protein context
+
+Protein context connects variant associations to protein architecture through Ensembl gene/transcript/translation IDs, canonical and protein-coding isoforms, exon-to-protein coordinate mapping, RefSeq/UniProt cross-references, Ensembl features, EBI Proteins features, and InterPro domains/families/motifs/regions. DataHub harmonizes source-specific API records into gene-level protein-context payloads. Production totals around 66.9k isoforms and more than 3.2 million protein feature annotations require production QA confirmation before release.
+
+## Gene profiles
+
+Gene profiles integrate nomenclature, gene summaries, protein cross-references, ontology/pathway membership, and curated source metadata from sources such as HGNC, NCBI Gene, UniProtKB, GOA/Gene Ontology, Reactome, Human Protein Atlas, and ClinGen where included. DataHub should preserve source IDs, source versions, access dates, cross-reference IDs, and source-specific licensing notes.
+
+## Clinical guidelines / guideline graph links
+
+Clinical guideline artifacts are generated primarily by HCG and HCG-KG. DataHub consumes release JSON, graph exports, or vector/serving artifacts when present and links genes to guideline snippets, recommendations, evidence classes, evidence levels, conditions, biomarkers, drugs/interventions, and source documents. Guideline snippets are context for interpretation and are not automated medical advice.
+
+## Drug-discovery / Drugs & Compounds layer
+
+The drugs and compounds layer uses Open Targets Platform GraphQL API v4 and licensed DrugBank v5.1.12 inputs where available. DataHub should preserve the GraphQL query, variables, access date, source field names, molecule source, molecule ID, target ID, source action type, indication, trial phase/status, source version, and source license. The raw DrugBank full database is license-restricted and must not be committed or archived unless redistribution permission is confirmed. The reported drug-layer total of 17,128 gene-drug records across 1,839 gene files and 1,454 unique molecule names requires production QA confirmation.
+
+## Expression and shared-architecture layers
+
+Expression payloads are imported from source-specific expression resources and existing HBP payloads where present. Shared genetic architecture is derived from association artifacts by comparing gene-level CVD and trait variant overlap. Both layers inherit redistribution constraints from their source data and must preserve source dataset, source version, input file, transformation, and HBP build version.