View Full Version : How Reliable Are Home DNA Ancestry Tests? Investigation Uses Triplets to Find Out

06-24-2017, 04:13 AM
Hadn't seen this story posted on here before but thought it might be of interest

The sisters were all 99 percent European but the test from 23andMe also showed some surprising differences.

Nicole was 11 percent French and German but Erica was 22.3 percent. Their sister Jaclyn was in the middle at 18 percent.

"I'm surprised," Nicole said. "I’m surprised because we came from the same egg and DNA. How are our ancestries different?"

Guerrero then met with the Maynard triplets and revealed the results for their “Family Tree DNA” tests. It showed they all had British Isles ancestry but the amount was different.

Erin Maynard was 59 percent, Mandy was 66 percent and Melissa was 70 percent.
The tests also showed that Mandy had six percent Scandinavian ancestry but her identical sisters showed none.

06-24-2017, 06:07 AM
I created a thread about it in Feb here (http://www.anthrogenica.com/showthread.php?9822-Inside-Edition-Puts-Ancestry-DNA-Kits-To-The-Test-Among-Sets-Of-Triplets)

Ann Turner had created a thread about it here (http://www.anthrogenica.com/showthread.php?9826-Different-results-for-ancestry-composition-when-uploading-the-same-file)

There is a third thread about the same segment of the show here (http://www.anthrogenica.com/showthread.php?9783-How-accurate-as-testing-kits-Triplets-Study)

C J Wyatt III
01-11-2018, 02:52 PM
I had a contact use this to say that DNA tests are not reliable, so he won't test. Of course all of this was about admixture/ethnicity and I would agree somewhat. However he is a potential 'first cousin, once removed" to a person who I have been working with and his would be a great autosomal DNA test result to have.

It appears to me that the "investigators" conveniently did not do "one-to-one" comparisons between the siblings. I would expect to see green across all chromosomes (using GEDmatch's chromosome browser).

Is everyone in agreement here?


Jack Wyatt

01-11-2018, 05:15 PM
DNA tests are highly reliable!
I have tested me at FTDNA, MyHeritage, Ancestry and LivingDNA - comparing the results between laboratories they match to 99,999% with just few SNPs being read differently (few, I haven't counted, but few as I see att gedmatch - those give half-match).
I have tested my mom at the same laboratories - same here, the results match 99,999%. My results and my mom's results match to 99,999%.
I have tested my uncle - his results match between laboratories and they match my mom and me at expected level.

DNA tests are reliable since they are "repeatable" and give consistent results even compared between several laboratories.

But the "results" of the tests are not always valid.
If we use the tests to research our relationships children-parents-grandparents-other relatives, then the results are absolutely valid. Well, I get clear "predictions" that my mom is my mom and that my uncle is my uncle (or possibly grand-parent if I remember correctly, but it is another story).

If we use the test to estimate ethnicity, then the results, or we should say the interpretation of the results, have low validity. The "tests for ethnicity" don't measure what we expect. This ethnicity hype drives me nuts! Think so: if we are descendants of 2000 or 20000 people who migrated, mixed, "mutated" during a hundred thousand (!) years, can we talk about ethnicity? What about migrations the last 2 thousand years? According to some interpretations I am 1% Native American... Really? Take those results with a pound of salt!

The tests (the technology there) are highly reliable and highly valid. The estimations of relationships (are we related) are reliable and valid, but probably are not so exact as we wish sometimes. No one can say that "here is your 6th cousine uncle trice removed and his grandgrandfather was adopted here and there for 150 years ago". The ethnicity estimates are not reliable and are not valid since "the results" (interpretations) could NOT be reproduced at different labs and, well, how we can measure ethnicity?

Hadn't seen this story posted on here before but thought it might be of interest


01-11-2018, 09:00 PM
Inside Edition is the "original" tabloid magazine, basically making pseudo-news for shock and controversy, not to inform. I think the bottom-line is there will always be parties criticizing any new, popular idea.
Those who have tried GEDMatch or MyHeritage with imported DNA data know that interpretation results can vary few percent between runs or samples from different companies, but I have not heard of huge margins. If there was such a huge problem as that segment implies, I'm sure several genetic genealogists would've already noticed, so by lack of articles from any reputable sources I'm tempted to think something else is wrong here.