Different sites match people differently. 23andMe has recently added very fine detail for regions as I reviewed inb my last post. For example:
When I go to Ancestry DNA — things are very different… no “Danish DNA” profiles..
Clearly, 23AndMe has much finer identification of regions which appears to be very good (relatively speaking).
With 23AndMe, I ended up being 51% ‘pure Danish (if you count Skane as Danish), 10% Schleswig-Holstein – Actually 61% Danish if you use maps from 1600. The 20% British is a complex issue because Middle-Jutland is where a lot of the people that settled in the English “Dane Law” came from. Look at the two maps below – one is my DNA in UK and one is DaneLaw. My UK areas according to 23AndMe matches DaneLaw — not surprising for a Dane ;-).
And last, MyHeritage DNA
Again overlapping areas explain the different attributions
Which Service or Both?
What are the main differences:
- 23andMe beats Ancestry.com for ancestors locations by several magnitudes. This is ideal for dealing with ‘missing paper records’ issues — you at least have some idea of where to look.
- Ancestory has 9 million+ Dna Samples. 23andMe has 5 million samples. So trying to find missing families especially for adopted kids — Ancestry beats 23AndMe, especially because of the intregration of Ancestry Trees.
I did both —
- 23andMe allowed your Ancestry Dna to be upload to them for a while. This was a special in 2018 (see this post) – unfortunately no longer available
” That means you can pay for a single test at AncestryDNA and copy the results to MyHeritage, Family Tree DNA, and Living DNA at no additional cost. This so-called transfer process won’t remove your DNA data (sometimes called a “kit”) from AncestryDNA, it’ll just add it to the other databases.” [src]
See this page for the various sites that you can transfer data to.
Ancestry Areas for their groupings
If you compare the regions identified by 23andMe, you see that Ancestry overlaps Denmark in all of the three regions they identified.
Life is even more complex!
My original AncestryDNA results estimates are below:
They revised their numbers and I became:
One of my daughters just got her results — and yes, ~ 50% of her DNA is from me, but her numbers are:
Let us write out some simple math: Mom + Dad = 200%, so kid % is typically (Mom% + Dad%) /2. This becomes
- Mom= Kids% * 2 – Dads%
So we have for Mom
- Swedish = 52* 2 – 39 = 65%
- Norwegian = 27 *2 – 18 = 36%
- Finnish = 10 * 2 – 0 = 20%
- German = 18 *2 – 0 = 36%
- England = 2 * 2 – 42 = -38%
Something is VERY WRONG — the answer is actually fairly simple:
- Different testing methods (Apples to Watermelons) using different reference sets -> different answers estimated.
There’s a second factor that is much harder to explain — overlapping patterns in DNA that are associated with multiple regions. One series may be associated with two places. If the mixture breaks the dna at the right spot, it is seen as English at another Finnish.
For finding DNA relatives— it is likely reliable
The reason is that we are matching precise sequences.