MyHeritage is happy to announce a significant enhancement in the interpretation of your DNA and ethnic makeup.
Technological advances have allowed our developers to enhance our algorithm, which calculates the origins of your DNA, making it even more sensitive to subtle differences between genetically close ethnicities. The amount of data used in this version and new algorithmic methods allow us to provide a more refined and accurate representation of your ethnic makeup, by enabling us to break down broader regions into even more refined ethnic regions and better represent existing ones.
Consequently, our Science team have expanded our reference populations (previously 42 ethnicities) to now include 79 different ethnicities from a wide variety of regions across the globe. Therefore, you may notice that you now have more ethnicities in your estimate, but with lower percentages of each population.
Likewise, if you are missing an ethnicity you previously had, and have a new ethnic region in your estimate, this is most likely because we removed that ethnicity from this version of the estimate and repartitioned it together with close ethnicities in a way that better reflects the genetic makeup of these regions combined. This allows us to provide you with a more precise representation of your ethnic makeup.
Examples of these changes include the region that was previously defined as “North West European”, which is now broken down and refined into “Dutch”, “Germanic”, “French”, and “Breton”. Another example is the region we referred to as “South Asian” in our previous estimate, now broken down into even more groups which we call: “Pakistani and Punjabi”, “South Asian”, “Bene Israel Jewish”, “Pashtun”, “Cochin Jewish”, and “Bengali”.
We note that even for ethnicities that remained unchanged, e.g. “East European”, you may still encounter small differences in estimated fractions, because the training of our new algorithm uses an extended sample size which gives a better and more even coverage of the genetic makeup of that ethnicity.
Beyond the differences in the reference panel used, the new estimate is based on new algorithms which, while better estimating a user’s ethnic decomposition on average, shows somewhat different error patterns compared to the previous version. While some common confusions between close ethnicities were improved, there are inevitably some specific confusions that are more typical of the new version.
To conclude, your Ethnicity Estimate — which is the result of an accurate and validated statistical algorithm — is still an estimate. Certain global populations exhibit similar DNA due to geographic proximity and the mingling of populations (See this FAQ for more details). We are always working on improving our algorithms and aim to give the most accurate results possible. Determining one’s ethnicity based on DNA is a relatively new field in genetics. Great advancements have been made in the past years, but there is much more to discover. We are constantly working on adding new ethnicities to enhance the Ethnicity Estimate based on new scientific discoveries.