Very different DNA results

[NGans]

Hi!
So, I tested at Myheritage and then uploaded it here. The results vary, to say the least.
Myheritage: 83,3 % North and western Europe
13,6 % Scandinavia
1,7 % Ireland, Scotland, Wales
1,4 % Ashkenazy Jew

FamilytreeDNA: 52% Scandinavia
29 % Eastern Europe
15 % West and central Europe
3 % Iberia
1% Ashkenazy Jew

I´m fully aware that this is not an exact science, and that different companies use different reference group, etc, but I can´t help wondering how it´s possible to get 29% Eastern Europe here and nothing at Myheritage...
Familytree is closer to the truth when it comes to the Scandinavian bit, my mother is "100%" Swedish. My father is a mixture of Danish, German, Austrian, Jewish and Swedish.
Also, when it comes to the Jewish heritage, since my father is approx. 12,5 % I would have expected that percentage to be a bit higher. At least I have lots and lots of Jewish DNA matches. Maybe someone more experienced in here have something wise to say about my results. Thank you in advance :-)

[John McCoy]

Estimating "ethnic origins", at least for Europe, where the underlying populations and genes have been thoroughly mixed through large-scale and frequent migrations throughout recorded history, is not an exact science. In fact, it may not even be a science.

In addition to the complicating factors you mentioned, there is also the problem of how to label the "reference groups" that are used by each vendor ("western Europe" and other labels may not mean the same thing to each vendor). Also, the way recombination and the random assortment of chromosomes works at each meiosis, you will not likely end up with exactly 12.5 percent of your DNA from each of your great-grandparents -- some will almost certainly be over-represented and others under-represented. At least one study estimated that by the time you go back 5 or 6 generations, at least one of those ancestors probably isn't represented AT ALL in your DNA.

Ethnic origins is a blunt instrument (like a tuba), it should not be expected to yield precise results.

[NGans]

Thanks for your reply!
So, I guess I didn´t inherit much DNA from my Jewish ancestor then. At first I started to doubt that he really was the father of my great grandmother, and that the small Ashkenazy percentage was from someone else further back that I didn't know about.

[Emona]

You can play with Chromosome Browser and check which segments of your dna are shared with your Jewish matches.
Likely more than 1%.

[clintonslayton76]

To bolster John McCoy's answer, Blaine Bettinger did a paper which suggests that the female share of atDNA recombines more actively than the male. This might help explain the wandering percentages of cMs and thus of theoretical "admixtures." I realize that you are concentrating on different expectations from your father, but it might help to know.
Here it is: Blaine T. Bettinger, "The recombination project: analyzing recombination frequencies using crowdsourced data." Preprint published on The Genetic Genealogist website but accessed through Scientific Papers section at (https://isogg.org/wiki/Recombination: accessed November 18, 2017)
I have found the Ashkenazi "admixtures" to be more notional than other researchers.

[John McCoy]

Yes, recombination in humans occurs more frequently in females than in males. That generalization very likely hides even more variation, because recombination is itself under genetic control and therefore it is expected that heritable natural variation in the frequency of recombination must also exist in humans, as it does in many other organisms where it has been possible to look for it. But it is extremely unlikely that you will ever know if your own family has a higher or lower rate of recombination than the general population, because you will never get a sufficiently large sample size to produce reliable statistics.

It may not be obvious, but the fact that, on average, recombination has the effect of transmitting the parental DNA in the form of several dozen "chunks" (even though correctly linked up into 22 autosomes), is the main reason that we don't end up with exactly, or even approximately 12.5 percent of our DNA from every one of our 8 great-grandparents. If recombination occurred much more frequently, the "chunks" would be much shorter and more numerous, and we would get closer to an even share from every great-grandparent, but the much shorter pieces of DNA would also be much harder to detect with the current matching algorithm.

It's all about the dance of the chromosomes during meiosis.

[NGans]

Thank you once again for interesting inputs!