AADR

AADR supplies release-owned human ancient-DNA metadata. The checked-in capture preserves a named release, persistent dataset identity, two annotation tables, and file identities so human samples can enter geographic publications without becoming an unversioned background layer.

In this repository, “human ancient DNA” means the governed metadata view used for identity, geography, chronology where supported, and contextual publication. It does not mean genotype analysis, imputation, ancestry inference, or a claim that the annotation panel is a complete census.

The current runtime consumes annotation metadata. It does not process AADR .geno, .ind, or .snp genotype files and does not claim population-genetic analysis.

Checked-In Release

Property	Governed value
requested release	v66
persistent dataset identity	doi:10.7910/DVN/FFIDCW
upstream Dataverse release	10.0
release timestamp	2026-04-13T04:33:11Z
annotation members	1240K and Human Origins (HO)
1240K annotation rows	23,250 records after the header
HO annotation rows	27,755 records after the header

The two row counts are table populations, not a unique-individual total. A person or genetic representation can occur across datasets, and identity must be resolved through the source identifiers rather than by adding counts.

flowchart LR
    Dataset["persistent AADR dataset"] --> Release["release v66"]
    Release --> K1240["1240K annotation table"]
    Release --> HO["HO annotation table"]
    K1240 --> Normalize["release-aware sample metadata"]
    HO --> Normalize
    Normalize --> Scope["world, regional, and country selection"]

Release Capture And Species Projection

The checked-in authority currently has two layers with different persistence:

Layer	Current repository state	Consequence
AADR release capture	data/aadr/v66/ contains the release manifest and both annotation members	release identity, file membership, checksums, and source rows are directly inspectable
species-owned raw route	data/adna/species/homo_sapiens/raw/aadr links to the AADR capture	human species ownership is visible without copying the release
species-owned normalized and review roots	the governed directories exist but contain no checked-in member or review artifacts	do not claim a persisted human normalization or review database from those roots
country publication bundles	checked-in country CSV, GeoJSON, Markdown, summary, and manifest outputs	publication rows are derived directly through the release-aware annotation runtime and must be audited against both bundle membership and the AADR source row

flowchart LR
    Capture["AADR v66 manifest and annotation rows"] --> Runtime["release-aware metadata projection"]
    Runtime --> Bundle["country bundle and manifest"]
    Capture --> RawRoute["Homo sapiens raw source route"]
    RawRoute -. no persisted members .-> Species["species normalized and review roots"]
    Bundle --> Audit["publication member to release row"]

This boundary is narrower than a fully materialized human evidence database. The country products are reproducible from the governed release and runtime, but a consumer must not cite the empty species-owned roots as proof that a separate normalized or reviewed population was checked in.

Publication Readiness Is Route-Specific

The current AADR lifecycle supports a release-to-country-publication route. It does not support a claim that every intermediate evidence stage has been persisted. Those are separate statements:

Question	Current answer	Authority
Can a source annotation row be inspected?	yes	named release member and checksum in release_manifest.json
Can a country feature be traced to its source row?	yes	country bundle membership plus the release-aware projection route
Is there a checked-in species-normalized human member database?	no	the species normalized root contains no governed members
Is there a checked-in species review population?	no	the species review root contains no governed review artifacts
Does a published feature prove genotype processing?	no	the runtime consumes annotation metadata only

The route can therefore reproduce a publication without pretending that a different, empty route has been completed. A future species-normalization workflow must create its own members, review decisions, and population receipt; it cannot retroactively treat country exports as those missing authorities.

Read An AADR Member

An annotation row can carry genetic, individual, skeletal, publication, repository, dating, locality, coordinate, data-type, and assessment fields. These fields do not all have the same authority or precision.

Claim	Evidence to retain	Boundary
sample identity	genetic and persistent identifiers plus release member	labels can represent genetic data instances rather than unique persons
publication lineage	DOI, publication abbreviation, and repository locator	first publication and the publication of this representation may differ
chronology	method, mean and deviation, full source wording, and date class	direct and contextual dates must remain distinguishable
locality	source locality, political entity, latitude, and longitude	map precision cannot exceed source metadata precision
data posture	pulldown strategy, data type, libraries, and assessment	metadata does not substitute for genotype processing

For a public claim, retain the AADR release and annotation member as well as the row identity. A sample label without release context cannot explain later coverage or metadata changes.

Audit A Human aDNA Feature

1. Record the publication scope, product version, layer, and feature identifier.
2. Resolve the feature to its AADR annotation member and source row.
3. Confirm the release identity in release_manifest.json; do not infer it from the map filename or sample label.
4. Read locality and chronology from the annotation fields at their reported precision, including method and source wording where present.
5. Distinguish the genetic-data instance from a unique person before counting or joining 1240K and HO rows.
6. Retain the product admission and geographic-scope decision separately from the source metadata.

This route separates three identities that are easy to conflate: the public map feature, the release-versioned annotation row, and the represented person or genetic data instance. A match at one level does not prove a match at the others.

Keep The Three Populations Separate

An AADR result may report three different populations. Each answers a different question and must retain its own denominator:

Population	Unit counted	Suitable claim
annotation member	release-versioned table row	rows present in one named AADR member
projected metadata row	runtime-admitted representation	records that passed the declared metadata and geography rules
country bundle feature	member in a governed publication bundle	features published for one declared country scope

Adding 1240K and HO row totals is not a person count. Likewise, counting country features is not a release total, because projection, geographic membership, and cross-member identity can change the population. Any aggregate must name the release, annotation member or members, admission rule, identity unit, and publication scope.

Release Changes Affect More Than Counts

Release difference	Re-evaluate
member file added, removed, or renamed	dataset membership and publication input identity
row identity or aliases changed	cross-member deduplication and stable feature lineage
locality or coordinate changed	geographic scope, point geometry, and distance relations
chronology fields changed	temporal posture and cross-family comparisons
publication or repository locator changed	citation lineage and recoverability
assessment or data-type field changed	metadata interpretation without implying genotype analysis

A release refresh is complete only when these semantic differences are reviewed. Equal row counts do not establish equal evidence.

Relationship To Animal aDNA

AADR is a release-oriented human metadata family. Animal aDNA is a project-and-literature recovery system whose publication depends on curated sample, locality, chronology, coordinate, and supporting-material evidence. They may share a map and both represent direct sample evidence, but neither inherits the other's recovery model, denominator, or fitness decision.

Spatial proximity between a human and animal point supports a qualified geographic comparison. It does not establish contemporaneity, association, or a shared archaeological context unless those dimensions are evaluated separately.

Review A Refresh As Population Reconciliation

An AADR refresh reconciles release members, annotation rows, projected Homo sapiens records, and product members as distinct populations. The receipt accounts for stable and changed identities, aliases across 1240K and HO, locality and chronology differences, projection refusals, country membership, and affected publications. It must explain both additions and removals. Equal release row counts do not establish equal people, equal projections, or equal country products.

Governing Surfaces

• data/aadr/v66/release_manifest.json governs persistent identity, upstream release, member files, sizes, and checksums;
• data/aadr/v66/1240k/v66.1240K.aadr.PUB.anno preserves the tracked 1240K annotation population; and
• data/aadr/v66/ho/v66.HO.aadr.PUB.anno preserves the tracked Human Origins annotation population;
• data/adna/species/homo_sapiens/raw/aadr exposes the release through the species-owned source boundary; and
• docs/report/countries/<country>/ contains the derived AADR publication bundles whose manifests govern country membership.

Continue to AADR exports for publication use, shared normalization for field lineage, and source comparison before combining AADR with another family.