Saturday, March 25, 2017

Ethnic Origins Predictions Affected by Random Autosomal Inheritance



Predictions of our ethnic origins are one of the biggest draws to get people to order a DNA testing kit. Almost all of these predictions are based on our autosomal DNA. That's the only DNA tested by Ancestry, 23andMe and MyHeritage. Family Tree DNA offers these predictions as part of its Family Finder test results. Marketing surveys suggest that more more than half of millennials who take genealogy DNA tests are doing it only or primarily to find out whether they should be wearing kilts or lederhosen to the next family gathering. 

In spite of the warnings of prominent genetic genealogists that these projections at present are one of the softest parts of the science of genetic genealogy, the advertising seen on television and elsewhere continues to emphasize predictions of ethnic origins. 

The Legal Genealogist, Judy G. Russell, is one of the most outspoken critics of the credibility of the predictions we are given. In an August 14, 2016 blog post "Those Percentages, if you must", she reiterated:
The Legal Genealogist says that the ethnicity estimates part of autosomal DNA tests are not a whole lot more than cocktail party conversation,
It just doesn’t seem to matter.
Really.
You can’t rely on DNA tests to give you exact percentages of your ethnic origins beyond the continental level (European, versus African, versus Asian).
 
Judy's conclusion is a valid one. Part of the problem is that this part of our science is still immature: 

  1. The population geneticists really can't tell us where all of our ancestors were living 500 years ago;
  2. Very few of us have pedigree charts that inform us about who all of our genealogical ancestors were in 1500 and where they resided; 
  3. Although atDNA tests generally sample 500,000 to 700,000 locations to make these ethnic predictions, this is a small percentage of our entire genomes; AND   
  4. Inheritance of our atDNA is random.
We all know that we inherited 50% of our atDNA from our fathers and 50% from our mothers. But how many of us know how much of our atDNA we inherited from our maternal grandfather? Would this make a difference in our predicted ethnicity? Of course it would.

Last month I published results from atDNA tests taken by two of my grandsons at MyHeritage. They are full siblings. Their other grandfather is primarily Ashkenazi. His Family Finder origins prediction is shown below. 




None of the other three grandparents show the presence of any Ashkenazi genetic component in their Family Finder results. My daughter-in-law was reported in her Family Finder results to be 42% Ashkenazi. [23andMe reported 48%.] Although I am not comparing apples to oranges in this post, I may be comparing different varieties of apples to each other by comparing testing results from three different companies. However, different family members have tested at different companies. These cross testing companies comparisons should not effect my main point in this post. If you have data that would dispute that, please share and we all will learn something.

My main point is that my grandson who appeared to inherit only 22% of his atDNA from his Ashkenazi grandfather in my previous post, was shown by Family Finder to have significantly less Ashkenazi ethnicity than did his full sibling who appears to have inherited 26% of his atDNA from that grandfather. 

This is an extremely small sample and the results may be distorted because results from different companies are being compared. However, I encourage you to try this at home if you have test results reflecting the difference between the atDNA inheritance patterns of sibling grandchildren. The result I report above is in accordance with what common sense would lead me to believe. If you have results that would support or refute this tentative hypothesis, I would love to see it. Maybe we all will learn something.


No comments:

Post a Comment