DNA Discrepancies

To create a customer’s ethnicity estimate, DNA companies–including AncestryDNA, 23&Me and MyHeritage DNA, to name a few–compare the customer’s DNA to a reference panel, or DNA from individuals of “known origin.” If a section of the customer’s DNA is most similar to the DNA tied to a reference panel of individuals from Sweden, for example, then that section of the customer’s DNA is “assigned” to Sweden. Identifying the best candidates for the reference panel is key to providing comprehensive ethnicity estimates; discrepancies arise between companies because each company has different DNA samples tied to their reference panels.

This week’s 52 Ancestors prompt is “surprise,” and instead of highlighting an ancestor’s story, I’m highlighting an unexpected turn in my own: my MyHeritage DNA results. AncestryDNA is the only company I have tested with, and I shared my updated ethnicity estimates back in September. As it turns out, 43 percent of my DNA links to Germanic Europe; 27 percent to Eastern Europe and Russia; 11 percent to England, Wales and Northwestern Europe; 9 percent to Ireland and Scotland; 5 percent to the Baltic States (primarily located in Estonia, Latvia and Lithuania); 3 percent to Sweden; and 2 percent to Norway. The most helpful result? Ancestry discovered a strong connection to Pomerania, providing me with a boundary-of-sorts for my research in Poland and Germany.

My AncestryDNA ethnicity estimates map

MyHeritage DNA allows customers to download their raw DNA data from other companies–including AncestryDNA–and upload it into their database to be analyzed; I thought I’d try it out, see if there’s anything new to learn or discover. (An aside: when I uploaded my raw data to MyHeritage, I was able to unlock their full results for free; now, customers must pay to unlock their results.) It took MyHeritage less than a day to create my ethnicity estimate and link my kit to potential cousins in their database, but there are discrepancies in the two companies’ estimates–and I don’t know what to make of these differences.

So here are my results: according to MyHeritage, 51.6 percent of my DNA links to Scandinavia; 35.6 percent to the Balkans; 8.2 percent to Eastern Europe; 3.5 percent to Ireland, Scotland and Wales; and 1.1 percent to the Baltic States. How do I account for some of these differences? Many individuals from Britain and Eastern Europe have significant “amounts” of Scandinavian DNA (due to travel between these regions); it’s possible that some of the “Germanic Europe,” “Eastern Europe” or “Northwestern Europe” DNA that Ancestry detected was placed into the Scandinavian category on MyHeritage. Likewise, the Balkan Peninsula is included in Ancestry’s Eastern European estimate; maybe MyHeritage is merely able to discern that difference.

But that much Scandinavian? That much Balkan? I have six great-great-grandparents who were born in Poland; one from Sweden; one from Germany; and eight who were born in the United States (all of German or English ancestry, based on the paper trail). Most of my Polish ancestors remained in the north–not in the Balkans–and I definitely don’t have any other Swedish ancestors in my family tree. Ethnicity estimates are fun–odd, but fun–and completely confusing, too. Maybe I’ll stick to painting my DNA and deciphering centimorgans and shared cousin matches (like that’s not confusing!)–or maybe I’ll try Family Tree DNA’s ethnicity estimates next. Until next time–