Ancestry DNA tests are pseudoscience. Not reliable to say the least.

Stuart Wright

AVForums Founder
Staff member
Just watched this video in which identical twins get different ancestry DNA results.
Am surprised to learn that ancestry DNA tests are only estimates and not reliable.
The ads present the tests as being scientific, but should actually be regarded as 'for entertainment purposes only'.
 

imightbewrong

Distinguished Member

DPinBucks

Distinguished Member
Just to be clear:

This is not pseudoscience. Astrology, for example, is pseudoscience.

This is perfectly good science used improperly for commercial gain. It was either inaccurate, in that the tests were not carried out properly; or inadequate, in that not enough bases were tested. Or both.
 

IronGiant

Moderator
I would add that the Quality Control may be suspect, and that somewhere along the line samples may be getting contaminated. This could even happen at the point of source given that the customer is sampling themselves.
Dear Client: we have analysed your saliva sample and can confirm you are 60% Aberdeen Angus, 20% Lettuce, 10% tomato, 10% human* and 5% Horse.

*Cross referencing our records we can predict that you were served lunch by a 15 year male who works in a Burger King in Huddersfield.
 

200p

Active Member
In theory, assuming no contamination, the reason for the differences might be because there was a delay between processing the 2nd twin's data. eg. there might have been hundreds or thousands more people added to their database to compare with in between the 1st test and the 2nd test. As well as maybe the software model was different or re-trained in between the 2 tests.

Also, identical twins don't necessarily have identical DNA.

Also, even if the DNA was identical, maybe even if there was there was no contamination, maybe the string of characters they stored to compare from the 1st twin was different to that from the 2nd?
 

IronGiant

Moderator
Also, identical twins don't necessarily have identical DNA.

Drilling down though this their DNA was 99.6% identical. Which sounds impressive, but then again our DNA is 98% identical to a chimpanzee.

Unless they meant the bases tested were 99.6% identical which suggests a wobble factor in the testing procedure. Either way, it's been known for years that these consumer tests are unreliable. Just like unreliable genealogy, it's easy to get things completely wrong. DNA tests like this should always be backed up by reliable genealogy evidence.
As such they are simply for entertainment only. My only surprise is that this isn't general knowledge amongst AVF people, given how often it has been discussed previously.
 

Stuart Wright

AVForums Founder
Staff member
DNA testing is not pseudoscience, but there's an argument to be made that Ancestry DNA testing, as presented in the 34.5% Norwegian type accuracy is not scientific as there is such a high % of uncertainty.
 

200p

Active Member
DNA testing is not pseudoscience, but there's an argument to be made that Ancestry DNA testing, as presented in the 34.5% Norwegian type accuracy is not scientific as there is such a high % of uncertainty.
Though on one of the sites (at around 9:20 in the video) they can get an option to select the confidence level (where the default may be 50%.) They can change the confidence to 90% and download the data and it says thing like "broadly European" instead of "x% Norwegian". They could/should show show the confidence (probability?) level in the normal part of the program (not just the data download bit) and allow it to be easily changed.

Though it still may not properly work for ancestry, since it's probably easy to get more data on current populations but hard to know (with a lot of data) which were in which places many years ago.
 

The latest video from AVForums

Podcast: Q Acoustics Q3030i, Humax Aura, Roku Streambar & WandaVision Reviews and more...
Top Bottom