MAXGEN schema · v1.13

MaxDNA

Genetic evidence, always private, never the genome itself.

Download MaxDNA v1.13 All versions https://opengenealogyai.org/schemas/maxgen/v1/dna.schema.json

Evidence that belongs to your relatives too

More than fifty million people have taken a consumer DNA test. A 47-centimorgan match (a centimorgan is the unit used to measure how much DNA two people share) does not prove a relationship on its own, but combined with a documentary trail to the same ancestor it raises confidence sharply. Genetic data is also information about every biological relative you have, including people who never consented. MaxDNA is designed to capture the genealogical value while making a leak structurally impossible.

What it stores, and what it cannot

MaxDNA stores match metadata: shared centimorgans, longest segment, segment count, optional chromosome segments for triangulation, haplogroups (deep ancestral lineages on the direct paternal or maternal line), ancestry composition, and kit metadata.

It cannot store a genome. raw_genotype_stored is a constant false that cannot be set otherwise. There is no field for a raw file.

The privacy architecture

Always tier2-private. The licence field has exactly one allowed value. A MaxDNA record can never appear in a public dataset, API response, or embedding, even for people long dead, because a living tester's data describes their ancestors too.
Hashed kit IDs. kit_id_hash holds an HMAC-SHA-256 hash (a one-way scramble that uses a secret key, so it cannot be reversed by guessing) of the kit number. Matches store match_kit_hash the same way. You can triangulate without ever seeing your cousin's kit number.
Consent is a required field. Values include explicit opt-in, guardian consent, public-dataset redistribution allowed, deceased before 2000, pending, and withdrawn. withdrawn starts a 30-day purge; after that the record is deleted outright, which is the one place MAXGEN deletes anything.

Methodology fields most tools forget

cm_map_version: different recombination maps (HapMap, deCODE and others) give centimorgan values that differ by up to 15% for the same segment. Recording which map was used makes matches comparable.
phasing_status: whether a shared segment is on the paternal or maternal chromosome, both, unphased, or unknown. Essential for serious triangulation.
endogamy_flag and endogamy_population: communities that married within themselves for generations (Ashkenazi, Acadian, Mennonite, island populations) break the standard relationship estimates. The flag tells downstream tools to recalibrate.

How it connects to the rest

person_id links the test to the MaxPerson who took it. Confirmed common-ancestor candidates on a match carry a confidence and an inference method (manual, tree intersection, triangulation, shared surname and geography). Those flow into MaxPerson's dna_evidence[], where they strengthen documentary relationship claims through the same noisy-OR combination used everywhere else.

Fields

Field	Type	Meaning
`dna_id` required	string	Globally unique identifier for this DNA assertion.
`schema_version` required	const "1.13"	MAXGEN version this DNA record conforms to.
`person_id` required	string	Foreign key to the MaxPerson this DNA test belongs to.
`test_type` required	enum	Type of DNA test performed.
`test_provider`	enum	Company or platform that performed or hosts the original test.
`test_date`	object	Approximate date the test was taken.
`kit_id_hash` required	string	SHA-256 hash of the original kit number.
`kit_source` required	string	Where this kit metadata came from (e.g.
`haplogroup_y`	string	Y-DNA haplogroup designation (paternal line), e.g.
`haplogroup_mt`	string	mtDNA haplogroup designation (maternal line), e.g.
`haplogroup_y_confidence`	number
`haplogroup_mt_confidence`	number
`ancestry_composition`	array of object	Ethnicity/admixture estimates from the test provider.
`endogamy_flag`	boolean	True if subject belongs to an endogamous population (Ashkenazi, Acadian, Mennonite, etc.) where standard cM thresholds are unreliable.
`endogamy_population`	string	Specific endogamous population if endogamy_flag is true.
`matches`	array of object	DNA matches with other kits.
`external_ids`	object	Cross-service kit aliases.
`is_living_flag` required	boolean	True if the DNA-test subject may be living.
`redistribution_license` required	const "tier2-private"	DNA data is ALWAYS tier2-private.
`raw_genotype_stored` required	const false	MAXGEN does not store raw genotype data (FASTQ, VCF, 23andMe raw export, etc.).
`external_raw_data_location`	string	Free-text note about where raw data is held (e.g.
`consent_status` required	enum	Legal basis for storing this DNA metadata.
`consented_at`	string
`consent_evidence_url`	string	Link to signed consent record (e.g.
`withdrawal_requested_at`	string	When the subject requested deletion.
`source_records`	array of string	MaxRecord IDs that document this DNA test (e.g.
`asserted_by` required	string	Agent ID or contributor ID that produced this DNA assertion.
`asserted_at` required	string
`notes`	string	Free-text annotations.

Example

Because MaxDNA records are private by definition, this site does not publish a real one. The validator offers a synthetic minimal record you can inspect and check.

Listen

Episode 4

MaxDNA — Genetic evidence without giving up privacy

23:23

Hashed kit IDs, consent, endogamy, phasing, and how DNA boosts documentary confidence.

Download MP3 · Study guide

← All seven schemas