Ethnicity Estimates

Everyone always wonders what’s going on when they get their ethnicity estimate from Ancestry, FamilyTreeDNA, or 23andMe. It hardly ever agrees with their paper trail. Take my paper trail, for instance. Here are the ethnicities of my eight great-grandparents—shown in a table and a pie chart.

Table

Ethnicity estimate Pie Chart

 

 

 

 

 

 

Pretty straightforward, right? Well, here is my ethnicity estimate from Ancestry.

My Ethnicity

 

It’s a lot more diverse than my paper trail, and where is my majority German ancestry? There are several explanations for this.

First, when Ancestry’s definition of, for example, Great Britain, is reviewed, we see that it includes Germany, as is shown below to the right. So my unexpected high level of Great Britain might be because it actually is including some of my Germany ancestry.

Great Britain Map

 

 

 

Second, Ancestry’s ethnicity estimates are an estimate of where your ancestors lived 500-1000 years ago. Most paper trails don’t go back this far—almost always not every line on our trees. A lot of intermingling of peoples happened during this time. For example, during the Viking Age, which was around 1,000 years ago, Scandinavian Norsemen settled in Norse Greenland, Newfoundland, and present-day Faroe Islands, Iceland, Normandy, Scotland, England, Ukraine, Ireland, Russia, Germany, and Anatolia. (https://en.wikipedia.org/wiki/Viking_Age) This might be an excellent explanation for my higher than expected Scandinavian percentage.

Third, Ancestry reports each region in percents. However, there are ranges associated with each percent. The range is provided by clicking on a specific ethnicity. For example, my 36% Great Britain is given the very broad range of 7% to 64%.

Great Britain Range

Fourth, if you consider that, on average, a generation takes 25 years, then Ancestry’s timeframe of 500-1000 years ago is from 20-40 generations ago. As is explained on the Types of DNA Page, autosomal DNA undergoes recombination when eggs and sperm are made. Recombination shuffles the DNA that each of our parents received from their parents. The shuffling results in mixtures of DNA from each grandparent. This shuffling of DNA means that you get different pieces of DNA that are used for ethnicity estimates.

Let’s focus on a single SNP, that is, a single point on your DNA. Let’s say it’s SNP #3,500, and this SNP is one of the places that’s tested for ethnicity. At each SNP location, everyone gets one SNP from their mom and one SNP from their dad. Thus, everyone has two SNPs at each location. Each SNP in this example has four possibilities: G, A, T, C.

Let’s say that at SNP #3,500, my brother and I inherited the following.

SNP #3,500
Brother Me
From Mom G Mat. Grandma

Norwegian

A Mat. Grandpa

German

From Dad T Pat. Grandpa

French

C Pat. Grandma

English

50% Norwegian 50% German
50% French 50% English

That is, my brother inherited the following.

  • G from our maternal grandma’s Norwegian DNA
  • T from our paternal grandpa’s French DNA.

In contrast, I inherited the following.

  • A from our maternal grandpa’s German DNA
  • C from our paternal grandma’s English DNA

If ethnicity were based only on SNP #3,500, our ethnicity estimate would be the

following.

  • Brother: 50% Norwegian and 50% French
  • Me: 50% German and 50% English

Looking just at this one SNP, my brother and I have very different ethnicity estimates. Of course, ethnicity estimates use more than just one SNP.

This example only goes back 2 generations to our grandparents. Think about all of the shuffling that took place over the last 20-40 generations, and you start to see another reason that your ethnicity estimate is just that–an estimate.

Below and to the right are my and my brother’s ethnicity estimates from Ancestry.

Brother Ethnicity

My Ethnicity

Look at all of the differences. First, there’s a huge 5-fold difference in our Great Britain percentages: 7% (my brother) and 36% (me).

Our Eastern European percentages are also very different: 11% (my brother) and 3% (me—in my Trace Region category).

Our trace regions are also very different. My brother has 2% European Jewish, whereas I have none. Under West Asia, we both have 2%. However, his is 2% Caucasus, and I have 2% Middle East. I also have <1% Asian South, and he has none. This probably should be disregarded, as ethnicities under 1% usually are considered background noise.

This example shows why testing siblings gives a fuller picture of one’s ethnicity. Your sibling might have an ethnicity that didn’t show up for you—or more or less of it. One person’s results don’t give as much information as two or more siblings’ results.

This isn’t to say that one’s ethnicity estimate is useless or just for fun. In contrast, they can be very useful in sorting out some people’s matches. For example, sometimes a person’s mother has a different ethnicity than the father. If that person’s matches have that maternal or paternal ethnicity, then this helps with an initial sorting of the matches into (1) probably paternal and (2) probably maternal. It’s one clue in determining who matches are.