The ancestry results produced by DNAPassport or BABYGlimpse are generated with HumanCode’s ancestry algorithm, which may produce different results compared to other ancestry products like National Geographic’s Geno 2.0 or 23andMe. Here's some information about why that may be the case.
Estimating Genetic Ancestry
Different populations around the world carry distinct genetic markers. By developing statistical tools for comparing these markers between global populations, researchers have formulated ways of estimating genetic ancestry for an individual. Ancestry estimates between products depend heavily on three components:
- The number of genetic markers a product measures to estimate your ancestry (more markers = more accurate probabilities)
- The number and type of populations that a product uses as references (more populations = more detailed insight into specific global regions)
- The statistical method used to calculate how similar your DNA is to the reference populations
HumanCode products identify approximately 17,000 genetic markers in your DNA and compare them to the markers found in individuals from 15 genetic reference populations. By comparing your DNA to the DNA found in different reference populations, we calculate your “closest genetic match” in DNAPassport or BABYGlimpse. Thus, the percentages displayed in our products for each ancestry group (e.g. “South European”) represent your genetic similarity to each of the reference populations.
More Data = More Detail
DNAPassport and BABYGlimpse, like all DNA-based ancestry products, are limited to the reference data sets that they have available for comparison. The reference populations that our products use come from the 1000Genomes data set, which does not currently include specific reference populations such as American Indian, Ashkenazi Jewish or West Mediterranean. As a result, DNAPassport will report the probabilities for users with these ancestries based on the next closest genetic ancestry match. For examples, individuals with Ashkenazi Jewish ancestry may see that their highest probability result is "South Europe" or another European populations.
There are some genetic ancestries that are trickier to distinguish from each other because of population mixture throughout history. In general, DNAPassport offers genetic ancestry insights based on 15 broad geographical regions, so some groups may overlap or match combinations of others (e.g. British and Northern European). However, other DNA-based ancestry products may have access to more detailed reference populations, and may measure more genetic markers, which allows them to perform a more in-depth ancestry analysis.
Depending on your genetic make-up, there may be some populations which our statistical model predicted to have a very low percentage (e.g. 5% or lower). If your genetic ancestry does not have a close match in our reference data sets, the model may have lower confidence in these predictions. If one of your predicted ancestries does not meet our stringent confidence and quality criteria, we include it in the "Undetermined" category while we continue to make improvements to our ancestry prediction pipeline. Note that these estimates get combined into a single "Undetermined" category, so it is possible to end up with a relatively high percentage of "Undetermined" results. However, HumanCode is focused on providing you the best experience possible, which is why we aim to show you high confidence predictions. We are constantly working to improve the resolution of our ancestry prediction and hope to include additional populations in future updates.
Exploring Ancestry Further
Here are two other articles that you mind find useful as you learn about ancestry predictions:
If you are interested in exploring your genetic ancestry in even greater detail, there are several ancestry-specific products on the Helix marketplace that are intended for more in-depth analysis. You can check them out here.
We hope this provided useful information! If you have questions along the way, we're here to help - just chat with us via the chat function (lower right of this screen) or email us at firstname.lastname@example.org.