Advanced Admixture Modeling

Multi-Way Fits

Use statistical analysis to test complex ancestry models. Determine which combinations of ancient and modern source populations best explain your genetic composition using NNLS statistical fitting.

Discover Your Ancestral Origins

Use advanced statistical analysis to test which combinations of ancient and modern source populations best explain your genetic composition. Each model shows fit quality (p-value) and component percentages.

Analysis Method
NNLS Fitting
Populations Tested
127 Models
Statistical Validation
p-values
Best Genetic Fit panel across k-way combinations at distance 0.13, listing eight source populations led by Azerbaijani (Shaki) 28.5 percent, Italian (Campania, Caserta) 27.0 percent and Turkish (Erzurum) 13.2 percent, with a stacked proportion bar.
Best multi-way fit for a raw DNA kit sample, with each source population's share and margin.

Advanced Admixture Analysis

1

Define Source Populations

Choose which populations to test as potential sources of your ancestry.

2

Statistical Modeling

Advanced algorithms test how well different mixture ratios explain your DNA.

3

Calculate Fit Scores

Each model receives a fit score indicating how well it matches your genetic data.

4

Identify Best Fit

Top-ranked models reveal the most likely ancestry composition and origins.

Understanding Your Results

Fit Score

  • •90%+ = Excellent fit
  • •80-90% = Very good fit
  • •70-80% = Good fit
  • •< 70% = Poor fit

Ancestry Percentages

  • •Show proportion from each source
  • •Add up to 100%
  • •Reflect optimal mixture ratio
  • •May indicate admixture events

Model Interpretation

  • •Top model most likely
  • •Similar high scores = uncertainty
  • •Single dominant model = clear ancestry
  • •Multiple good models = complex history

Check the fit distance before you trust any percentages: the guide on how accurate an admixture calculator is explains what the distance measures and what counts as a good fit. To model yourself against ancient groups rather than modern ones, read the ancient DNA admixture calculator guide, and for when a formal test is the better tool, see how qpAdm differs from G25.

Frequently Asked Questions

Yes. Multi-Way Fits runs on your G25 coordinates. It models them as a mix of 2 to 8 source populations and keeps the combinations with the lowest fit distance, so you can see which mixture reproduces your coordinates most closely.
Models can use the 5,453 modern references, or the ancient samples of any one of ten periods from the Paleolithic to the Early Modern era.
A model is only as good as the references available. A low fit distance means the mixture reproduces your coordinates closely; it does not prove those populations were your ancestors. Complex ancestry often shows several competing models.
Yes. If you already have them you can import them. If not, start from your 23andMe, AncestryDNA, MyHeritage or FamilyTreeDNA raw data file: every paid plan includes a raw DNA conversion.
Yes. When reference populations are added or methods improve, running a model again can refine the result.
Multi-Way Fits is part of every paid plan. The Population Encyclopedia is free to browse with a free account; analysing your own DNA needs a paid plan.

Test Your Ancestry Models

Available on any paid plan. Explore complex admixture and ancestry composition.

Related tools

More from Advanced Modelling, and other tools that work from the same upload.

Hunter-Gatherer & Farmer Split

Break your ancestry into the deep components that shaped all modern populations: WHG, Anatolian Farmer, Steppe and more.

Archaic Admixture

Find out how much Neanderthal and Denisovan DNA you carry and what it means.

Unsupervised Analysis

Let machine learning find hidden genetic structure in your DNA without preconceptions.

Ghost Ancestry Mode

Detect "missing" ancestry from extinct or unsampled populations hiding in the gaps of your DNA.

See every tool