Commercial tests often report 0% Indigenous for people whose Native ancestry sits six or more generations back. We go to the raw file, the chromosome segments and the reference genomes, and report what the data actually supports.
Morgan Heights Studio has moved away from game design to work full time on Indigenous genomics. The work is more interesting, and heavier, than expected.
A written evidence report on whether your DNA carries Indigenous American ancestry, where it sits on your chromosomes, roughly how much, and which lineages it most plausibly came through.
Rare Native genomes have sat in lab datasets their entire existence, out of reach of the families they belong to. We rebuild them from the Allen Ancient DNA Resource as raw DNA files and port them to GEDmatch where people can compare.
The groundwork behind every report: reference-kit surveys, ancestry-informative marker panels built from public genome projects, and haplotype tests calibrated against European controls.
Independent methods converging on the same chromosome locations, with every number shown as measured.
Segment evidence and genome-wide averages measure different things. A few real Native segments can exist in a genome whose average is too low for averaging methods to resolve. A low panel number never cancels segment evidence, and it is never hidden either.
Specific stretches of DNA shared with Indigenous reference kits on GEDmatch, the most direct evidence available.
Answers: is it there, and whereMatches who share the same segment with you and with each other, showing a single inherited haplotype from one ancestor. Overlapping is not the same as triangulated; a triangle needs all three pairs.
Answers: one inherited stretch, or coincidenceA relative run against the same reference kits with identical settings. Zero matches where you have many rules out a pile-up region that matches everyone.
Answers: is the spot itself the problemEurogenes K13, MDLP World-22, Dodecad World-9 and HarappaWorld chromosome painting: genome-wide percentages and per-chromosome spikes.
Answers: how much, and which chromosomesA genome-wide estimate from allele frequencies at about 2,360 ancestry-informative markers that separate Native genomes from European ones, with its detection limit reported alongside.
Answers: roughly how muchNo public Eastern Woodlands reference genomes exist, and which kits matched does not identify a nation. Tribal attribution rests on documentary genealogy only.
A segment shared with a reference donor shows shared Native ancestry. DNA alone can't trace it to a specific recent ancestor, so we don't use it to place you in a particular family tree. Kinship is defined by the people, not by a match list.
Reference kits within one population are often related to each other. A spot matched by Pima, Athabaskan and ancient kits together counts for more than five kits from one family.
Every report lists cM, SNP count, density and kit count as measured, then labels interpretation as interpretation.
The Allen Ancient DNA Resource holds genotypes for thousands of published ancient and present-day individuals. Many of its rare Native American samples have sat in lab datasets their entire existence, never on the comparison sites people actually use.
Identify Native American and Beringian samples in AADR with enough coverage of consumer-chip positions to be worth comparing.
Pull each individual's genotypes and align positions and alleles to the build and strand consumer files use.
Write a raw DNA file in a format GEDmatch accepts, recording the SNP count that actually made it in.
Upload to GEDmatch as a research kit, labeled with the sample ID and source publication, so anyone can run a one-to-one against it.
Ancient genomes usually cover only part of what a consumer chip reads, so a match against one is reported with its SNP count and density like any other segment. Kits ported from published data are credited to the original study and the AADR compilers.
Most people who come to us were told by another test that they had no Native ancestry. The report either confirms that or sets the record straight, with the numbers to back it.
Third-party GEDmatch matches appear in reports by kit number only, relatives by initials, and no email addresses are included.
Your raw file goes through the AIM scanner, and the four calculator results are combined into per-model and mean Amerindian + Siberian totals with spike chromosomes marked.
GEDmatch 3D Chromosome Browser runs against the full reference set at 3 cM, with default one-to-ones on any kit that lands on a hotspot to see if the segment holds.
A relative is run as negative control, then Tier 1 Segment Triangulation and Segment Search on the top hotspot to find who shares it with you and each other.
Each hotspot is compared with the HGDP genomes, using 155 Europeans as the control, to see whether your sharing with Native genomes stands out.
A PDF with methods, a genome-wide results table, ranked hotspots, limitations and plain conclusions. Findings first, caveats referenced, nothing rounded away.
GEDmatch Indigenous reference kits used in the segment survey, grouped by population. Two pairs are the same genome uploaded twice and are counted once in results.
| Population | Region | Kits | |
|---|---|---|---|
| Cree | Canada | 8 | |
| Dene / Athabaskan | Canada | 12 | |
| Algonquin | Canada | 6 | |
| Ojibwe / Métis | Canada | 8 | |
| Sioux | USA | 1 | |
| Chippewa | USA | 1 | |
| Navajo | USA | 2 | |
| Pima | USA | 8 | |
| Cherokee (Willerslev) | USA | 4 | |
| South American | Colombia | 2 | |
| Guarani | Paraguay | 1 | |
| Ancient | Canada / USA | 4 | |
| Total listed | 57 | 55 distinct genomes | |
Mi'kmaq 1492, Kennewick Man, Clovis Montana (Anzick) and the Anzick research upload, which is the same genome.
There are no public Eastern Woodlands reference genomes. Results may under-read Eastern Native ancestry, and reports say so.
Several reference sets share hundreds of cM among themselves, so same-population backing is partly family sharing.
Curated genotypes of published ancient and present-day individuals. The source for ported ancient kits.
Mallick S, Micco A, Mah M, et al. (2024). Scientific Data 11: 182. Reich Lab, Harvard Medical School.929 genomes from diverse populations. Pima and Maya refine the marker panel; Europeans serve as the haplotype-test control.
Bergström A, McCarthy SA, Hui R, et al. (2020). Science 367: eaay5012.Public Mixe, Mixtec and Zapotec genomes blended into Native allele frequencies. Restricted-access samples are not used.
Mallick S, Li H, Lipson M, et al. (2016). Nature 538: 201–206.Base allele frequencies for the marker panel, plus 116 Native markers from the Galanter panel built for admixture across the Americas.
1000 Genomes Consortium (2015). Nature 526: 68–74. Galanter JM et al. (2012). PLoS Genet 8: e1002554.Morgan Heights Studio is the research imprint of Lone Star Tsalagi, rooted in Cherokee heritage. It started as an indie game and web studio. Building tools to read Native ancestry out of raw DNA turned into the work itself.
The studio now works full time on Native American DNA testing, recovering ancient DNA from lab datasets, and the population research that holds both up.
"I did not think I'd find the work this interesting or this heavy, yet here I am."
Sèitheach, founderTell us what you're after: an evidence report on your own DNA, a sample you want recovered, or research you'd like to work on together. We'll reply with exactly what to send and how.