Strongly varying percentages between 23andme and AncestryDNA

[Götterfunke]

Hello everyone,

There is quite the big difference in terms of percentages between my 23andme and AncestryDNA results (especially considering the Greek part): https://puu.sh/GgPhz/d109cb47ee.png

Note that the 23andme result is already phased with my 100% German father (with Dutch influences) ... before the phasing, the result was closer to 35% Greek, opposed to the 3% Greece, Turkey & Albania on AncestryDNA.

I wonder, did anyone experience something similar, or even stronger variations?

Also, which test was more accurate for you?

Don't hesitate to share your opinions!

[marco]

Ancestry dna was the more accurate of the two for me to be honest it got all my percentages correct, it’s the closest to my global25

[Leto]

Deine 23andme-Ergebnisse sind nicht besonders arisch ;)
(I understand German but my English is better)

3 vs 30 percent Greek is a huge-ass difference if you ask me. Are you really only 1/16 Greek?

[Götterfunke]

Deine 23andme-Ergebnisse sind nicht besonders arisch ;)
(I understand German but my English is better)

3 vs 30 percent Greek is a huge-ass difference if you ask me. Are you really only 1/16 Greek?

It indeed is-and the difference was even bigger than that before I phased with my father.

My father has no Greek genes whatsoever, and both of my mothers parents are German, too.

I am not sure whether or not I'm 1/16 Greek at all. It could very well be 1/32 or 1/8 as well.

I still have no information on who exactly was the last 100% Greek family member, but my great-grandmother (who didn't look typically Greek at all and was classified as borreby and typically West German on this forum) had a last name ending in -is, so it is likely that one of her ancestors was completely Greek, possibly her father or paternal grandfather.

It is very likely that he / she was from Peloponnese though, as this is the only region where I have strong evidence (according to 23andme) of recent ancestry. Most of my Greek cousins are also from there.

[Leto]

It indeed is-and the difference was even bigger than that before I phased with my father.

My father has no Greek genes whatsoever, and both of my mothers parents are German, too.

I am not sure whether or not I'm 1/16 Greek at all. It could very well be 1/32 or 1/8 as well.

I still have no information on who exactly was the last 100% Greek family member, but my great-grandmother (who didn't look typically Greek at all and was classified as borreby and typically West German on this forum) had a last name ending in -is, so it is likely that one of her ancestors was completely Greek, possibly her father or paternal grandfather.
Does GEDmatch manage to detect your alleged Southern European?

[Götterfunke]

Does GEDmatch manage to detect your alleged Southern European?

I usually cluster with Central Europeans, somewhere between East and West Germans, and on this forum also with some Hungarians.

These are my K-13 results: https://puu.sh/GgS1D/a51f149399.png

[Leto]

I usually cluster with Central Europeans, somewhere between East and West Germans, and on this forum also with some Hungarians.

These are my K-13 results: https://puu.sh/GgS1D/a51f149399.png
Yep, the Southern Euro ancestry is evident.

Distance to: Götterfunke
4.61432552 Austrian
6.94068440 Swiss_German
7.51240847 Hungarian_Transdanubia
7.59742720 East_German
8.20950668 Swiss_French
8.30896932 Hungarian_Budapest
8.31737338 Croat_Croatia_Gorski_Kotar
8.39751154 West_German
9.09800345 Hungarian
9.49952704 Hungarian_Szekely

Target: Götterfunke
Distance: 2.3481% / 2.34806442 | ADC: 0.5x
64.0 Austrian
28.0 Swiss_German
5.0 Pomak_Kavala
3.0 Turk_Sliven-Alevi

Target: Götterfunke
Distance: 1.4782% / 1.47820608 | ADC: 0.25x
59.8 Austrian
30.8 Swiss_German
6.2 Pomak_Kavala
2.2 Tabassaran
0.6 Karitiana
0.4 Papuan

[Götterfunke]

Yep, the Southern Euro ancestry is evident.

Distance to: Götterfunke
4.61432552 Austrian
6.94068440 Swiss_German
7.51240847 Hungarian_Transdanubia
7.59742720 East_German
8.20950668 Swiss_French
8.30896932 Hungarian_Budapest
8.31737338 Croat_Croatia_Gorski_Kotar
8.39751154 West_German
9.09800345 Hungarian
9.49952704 Hungarian_Szekely

Target: Götterfunke
Distance: 2.3481% / 2.34806442 | ADC: 0.5x
64.0 Austrian
28.0 Swiss_German
5.0 Pomak_Kavala
3.0 Turk_Sliven-Alevi

Target: Götterfunke
Distance: 1.4782% / 1.47820608 | ADC: 0.25x
59.8 Austrian
30.8 Swiss_German
6.2 Pomak_Kavala
2.2 Tabassaran
0.6 Karitiana
0.4 Papuan

Apparently close to 6-8%? 6.25% would equal one Greek great-great grandparent, or 1/16. That's what I always suspected.

[Teutonski]

It indeed is-and the difference was even bigger than that before I phased with my father.

My father has no Greek genes whatsoever, and both of my mothers parents are German, too.

I am not sure whether or not I'm 1/16 Greek at all. It could very well be 1/32 or 1/8 as well.

I still have no information on who exactly was the last 100% Greek family member, but my great-grandmother (who didn't look typically Greek at all and was classified as borreby and typically West German on this forum) had a last name ending in -is, so it is likely that one of her ancestors was completely Greek, possibly her father or paternal grandfather.

It is very likely that he / she was from Peloponnese though, as this is the only region where I have strong evidence (according to 23andme) of recent ancestry. Most of my Greek cousins are also from there.

Antworte doch auf deutsch?

[Leto]

Apparently close to 6-8%? 6.25% would equal one Greek great-great grandparent, or 1/16. That's what I always suspected.
Well, the thing is that both Austria and Switzerland are the Southernmost parts of German-speaking Europe. So some of the S. Euro ancestry may have been sucked into that.
Honestly I don't know what else to suggest.

[Götterfunke]

Well, the thing is that both Austria and Switzerland are the Southernmost parts of German-speaking Europe. So some of the S. Euro ancestry may have been sucked into that.
Honestly I don't know what else to suggest.

Interesting ... you already helped a lot. Thanks for the suggestions.

[Teutonski]

Leto, um wat wetten wir dat der Jung nicht in DE lebt oder deutscher ist?

[Leto]

Leto, um wat wetten wir dat der Jung nicht in DE lebt oder deutscher ist?
Glaubst du dass er eine Sockenpuppe ist oder sein kann? :D

[Ion Basescul]

I usually cluster with Central Europeans, somewhere between East and West Germans, and on this forum also with some Hungarians.

These are my K-13 results: https://puu.sh/GgS1D/a51f149399.png

AncestryDNA is definitely trash in their ethnicity estimate.

Target: Gotterfunke
Distance: 1.7173% / 1.71728620
37.4 German_West
25.6 Greek_Peloponnese
12.4 German_east
11.8 North-Dutch
11.6 Polish
1.2 Swedish
0.0 GR_Macedonia
0.0 GR_Peloponese
0.0 Greek_Central-Macedonia
0.0 Greek_Eastern-Macedonia
0.0 Greek_Eastern-Thrace
0.0 Greek_Macedonia_Thrace
0.0 Greek_Northern-Thrace
0.0 Greek_Thessaloniki
0.0 Greek_Thessaly
0.0 Greek_Western-Macedonia
0.0 Greek_Western-Thrace
0.0 South_Dutch

I suggest reading about precision and recall rates. With a precision of 93%, when 23andme assigns Greek & Balkan then there is a 93% chance for them to be right. This category however has a recall of 60%, meaning that they might struggle to assign all the real Greek & Balkan admixture there. However, both rates are significantly better than for French & German for example, where the numbers are 83% and 28%.

https://www.23andme.com/en-gb/ancestry-composition-guide/

[Götterfunke]

AncestryDNA is definitely trash in their ethnicity estimate.

Target: Gotterfunke
Distance: 1.7173% / 1.71728620
37.4 German_West
25.6 Greek_Peloponnese
12.4 German_east
11.8 North-Dutch
11.6 Polish
1.2 Swedish
0.0 GR_Macedonia
0.0 GR_Peloponese
0.0 Greek_Central-Macedonia
0.0 Greek_Eastern-Macedonia
0.0 Greek_Eastern-Thrace
0.0 Greek_Macedonia_Thrace
0.0 Greek_Northern-Thrace
0.0 Greek_Thessaloniki
0.0 Greek_Thessaly
0.0 Greek_Western-Macedonia
0.0 Greek_Western-Thrace
0.0 South_Dutch

I suggest reading about precision and recall rates. With a precision of 93%, when 23andme assigns Greek & Balkan then there is a 93% chance for them to be right. This category however has a recall of 60%, meaning that they might struggle to assign all the real Greek & Balkan admixture there. However, both rates are significantly better than for French & German for example, where the numbers are 83% and 28%.

https://www.23andme.com/en-gb/ancestry-composition-guide/

Highly interesting, thanks for sharing.

May I know how you calculated these results and where you got the samples from? I'm kind of new to the whole gedmatch and vahaduo stuff.

[Teutonski]

Glaubst du dass er eine Sockenpuppe ist oder sein kann? :D

100%

[Ion Basescul]

Highly interesting, thanks for sharing.

May I know how you calculated these results and where you got the samples from? I'm kind of new to the whole gedmatch and vahaduo stuff.

Yep, I got the samples from here (http://vahaduo.genetics.ovh/k13-vahaduo.htm) and limited the model to only those that seemed relevant for your ancestry.
To run yourself against the model, you need to follow the instructions from here (http://vahaduo.genetics.ovh/howtouse.htm).

[ShieldWolf]

Leto, um wat wetten wir dat der Jung nicht in DE lebt oder deutscher ist?


Glaubst du dass er eine Sockenpuppe ist oder sein kann? :D

Ich denke er ist ein Amerikaner.

[Luso]

This is a pretty typical issue I think.

I haven't done 23andMe but based on what my parents (who have done both tests) got... my results are going to say I am around 80% Iberian+around 6% Italian, and 10-12% northwestern European. While my ancestry results show 69% Portugal, 14% Spain, 7% France, 6% Irish or Scottish, 2% Greek/Balkans, and 2% Sardinia.

Personally I can only go based on my G25 and gedmatch in terms of accuracy... I will say 23andMe seems more reasonable? But for me (and other Iberians) both are reasonably close at distinguishing I'm Iberian. The only thing that 23andMe includes and ancestry doesn't is possible bits of North African (1% max), while ancestrydna says I'm 100% European.

[Luke35]

Your K13 certainly deviates notably from German results I am familiar with, like my mother's.

I know on paper you can't figure out where all that Greek/southern European comes from, but I can tell you 23andMe was spot on with my ancestry prediction. I trust that report more than AncestryDNA.

[Lucas]

Hello everyone,

There is quite the big difference in terms of percentages between my 23andme and AncestryDNA results (especially considering the Greek part): https://puu.sh/GgPhz/d109cb47ee.png

Note that the 23andme result is already phased with my 100% German father (with Dutch influences) ... before the phasing, the result was closer to 35% Greek, opposed to the 3% Greece, Turkey & Albania on AncestryDNA.

I wonder, did anyone experience something similar, or even stronger variations?

Also, which test was more accurate for you?

Don't hesitate to share your opinions!

This means nothing. Chech if both raw files have very different results in the same calculators on Gedmatch.

[Götterfunke]

This means nothing. Chech if both raw files have very different results in the same calculators on Gedmatch.

The differences are minimal.

[Götterfunke]

This means nothing. Chech if both raw files have very different results in the same calculators on Gedmatch.

The result on the left is AncestryDNA, the one on the right 23andme:

https://puu.sh/GhvH2/a918aa8de1.png

[Lucas]

The result on the left is AncestryDNA, the one on the right 23andme:

https://puu.sh/GhvH2/a918aa8de1.png

Yes as usual. For me it is best comparison if raw files are different or not in substantial degree.

[Götterfunke]

Yes as usual. For me it is best comparison if raw files are different or not in substantial degree.

According to the K13 admixture, which result would you consider more accurate, 23andme or AncestryDNA?

[Ion Basescul]

According to the K13 admixture, which result would you consider more accurate, 23andme or AncestryDNA?

There isn't such a thing as more accurate according to K13 admixture. One is more accurate than the other because it works with more DNA data from you. In your case, this is AncestryDNA.
Run the calculator on GEDmatch, and in the screen that's showing the pie chart of admixture, look down and you will see the number of SNPs used in calculation.
Higher = more data from you = more objective

[Götterfunke]

There isn't such a thing as more accurate according to K13 admixture. One is more accurate than the other because it works with more DNA data from you. In your case, this is AncestryDNA.
Run the calculator on GEDmatch, and in the screen that's showing the pie chart of admixture, look down and you will see the number of SNPs used in calculation.
Higher = more data from you = more objective

So AncestryDNA raw data is objectively more precise than the raw data from 23andme?

[Ion Basescul]

So AncestryDNA raw data is objectively more precise than the raw data from 23andme?

Yep, but the analysis on the commercial website is obviously better for 23andme.

[Ion Basescul]

AncestryDNA is definitely trash in their ethnicity estimate.

Target: Gotterfunke
Distance: 1.7173% / 1.71728620
37.4 German_West
25.6 Greek_Peloponnese
12.4 German_east
11.8 North-Dutch
11.6 Polish
1.2 Swedish
0.0 GR_Macedonia
0.0 GR_Peloponese
0.0 Greek_Central-Macedonia
0.0 Greek_Eastern-Macedonia
0.0 Greek_Eastern-Thrace
0.0 Greek_Macedonia_Thrace
0.0 Greek_Northern-Thrace
0.0 Greek_Thessaloniki
0.0 Greek_Thessaly
0.0 Greek_Western-Macedonia
0.0 Greek_Western-Thrace
0.0 South_Dutch

I suggest reading about precision and recall rates. With a precision of 93%, when 23andme assigns Greek & Balkan then there is a 93% chance for them to be right. This category however has a recall of 60%, meaning that they might struggle to assign all the real Greek & Balkan admixture there. However, both rates are significantly better than for French & German for example, where the numbers are 83% and 28%.

https://www.23andme.com/en-gb/ancestry-composition-guide/


Same model with AncestryDNA

Target: Gotterfunke_AncestryDNA
Distance: 1.4236% / 1.42355780
41.6 German_east
37.0 South_Dutch
18.2 Greek_Eastern-Macedonia
3.2 Greek_Thessaly
0.0 German_West
0.0 Greek_Peloponnese
0.0 North-Dutch
0.0 Polish
0.0 Swedish
0.0 GR_Macedonia
0.0 GR_Peloponese
0.0 Greek_Central-Macedonia
0.0 Greek_Eastern-Thrace
0.0 Greek_Macedonia_Thrace
0.0 Greek_Northern-Thrace
0.0 Greek_Thessaloniki
0.0 Greek_Western-Macedonia
0.0 Greek_Western-Thrace

Removed all Greeks apart from Peloponnese, since that's where you are listing your ancestry from.

Target: Gotterfunke_AncestryDNA
Distance: 1.5223% / 1.52229709
42.8 German_east
37.8 South_Dutch
19.4 Greek_Peloponnese
0.0 German_West
0.0 North-Dutch
0.0 Polish
0.0 Swedish
0.0 GR_Peloponese


Both result in better distances than with the 23andme data.

[Alenka]

What amount of SNPs did each have when you ran K13?

[Götterfunke]

What amount of SNPs did each have when you ran K13?

171857 for AncestryDNA, 58596 for 23andme.

[Harald Meller]

If you are 1/4 from Upper Silesia this explains the high percentage of Eastern European.

I already showed you ethno-linguistic data for the whole Province of Upper Silesia here:

https://en.wikipedia.org/wiki/Province_of_Upper_Silesia#Historical_population

Here you can see the ratio of Polish:German ("others" were ignored) by county by year :

https://i.imgur.com/BEp4SOJ.png

The figures are from this book: https://archive.org/details/diepoleninobersc00webeuoft

PS:

It would be very interesting to see the DNA results of some German with ancestry from Grottkau-Neisse area.

[Harald Meller]

If your Upper Silesian grandparent is alive, you might want to test her / him too.

Then you can phase your DNA with her / his to see if your East Euro % changes.

[Leto]

@Harald Peterski, hello, brat! ;)

[Götterfunke]

If your Upper Silesian grandparent is alive, you might want to test her / him too.

Then you can phase your DNA with her / his to see if your East Euro % changes.

I had the chance to attend her funeral three years ago. I think she wouldn't have been up for it either way, as her DNA-test would have revealed something she might have known-one of her daughters was not from my grandfather. But that's a different story.

I wonder what control groups they have used ... there are still Germans living in Poland, some simply changed their name and learned Polish to avoid persecution, others have mixed with the Poles.

Considering this, is it really beyond a doubt sure that the 26% Eastern European or the 12% Polish are actually "slavic" ancestry, and not East German (Silesian)?

[Harald Meller]

there are still Germans living in Poland, some simply changed their name and learned Polish to avoid persecution, others have mixed with the Poles.

Yes but it is said that most of these "Germans" are of Slavic/Polish origin, because after WW2 Polish Communist authorities kind of did a "racial selection" similar to what the Nazis had done during WW2. So they only allowed to stay those who were judged as "of Slavic origin and suitable for Re-Polonization".

For example, this couple identify as "Germans" but look at their phenotypes (very Slavic faces...) and name (Golebiowski...):


https://www.youtube.com/watch?v=rRp-dVzBGtg

[Götterfunke]

Yes but it is said that most of these "Germans" are of Slavic/Polish origin, because after WW2 Polish Communist authorities kind of did a "racial selection" similar to what the Nazis had done during WW2. So they only allowed to stay those who were judged as "of Slavic origin and suitable for Re-Polonization".

For example, this couple identify as "Germans" but look at their phenotypes (very Slavic faces...) and name (Golebiowski...):


https://www.youtube.com/watch?v=rRp-dVzBGtg

Last names are pretty irrelevant considering the fact that they were forced to change them, but you're right about the looks, they look somewhat Polish.

Still, not really helpful in my case. I guess I can only continue to speculate. It is theoretically possible that my great-grandmother was Polish, albeit I doubt it as my great-grandfather certainly wasn't.

[Harald Meller]

are actually "slavic" ancestry, and not East German (Silesian)?

Silesian Germans = a simple mixture of West Germans and Slavs, check this book:

https://archive.org/details/bub_gb_CqpCAAAAIAAJ (Karl Weinhold, "Die Verbreitung und die Herkunft der Deutschen in Schlesien")

[Harald Meller]

but you're right about the looks, they look somewhat Polish

Paradoxically one of the most genetically Germanic parts of modern Poland is south-eastern Poland, this region:

https://en.wikipedia.org/wiki/Podkarpackie_Voivodeship

A lot of Germans settled there in Medieval timees but were completely assimilated/absorbed into Polish ethnicity.

[Götterfunke]

Silesian Germans = a simple mixture of West Germans and Slavs, check this book:

https://archive.org/details/bub_gb_CqpCAAAAIAAJ (Karl Weinhold, "Die Verbreitung und die Herkunft der Deutschen in Schlesien")

It's questionable how much ancient ancestry matters for dna-tests though as 23andme only test ancestry of the recent 200 years or so-and their control groups are considerably younger.

[Harald Meller]

It's questionable how much ancient ancestry matters though as 23andme only test ancestry of the recent 200 years or so-and their control groups are considerably younger.

Really? I don't believe that. Explain why Mexicans do not score 100% Mexico in 23andMe, but rather 50% Spain + 50% Native American.

After all, in the last 200 years they are 100% from Mexico, so they should not score like a mix of two things?

[Götterfunke]

Really? I don't believe that. Explain why Mexicans do not score 100% Mexico in 23andMe, but rather 50% Spain + 50% Native American.

After all, in the last 200 years they are 100% from Mexico, so they should not score like a mix of two things?

Nevermind, the 200 years thing is false, this is what they say about their control groups:

"Customers comprise the lion's share of the reference datasets used by Ancestry Composition. When a 23andMe research participant tells us that they have four grandparents all born in the same country — and the country isn't a colonial nation like the US, Canada, or Australia — that person becomes a candidate for inclusion in the reference data. We filter out all but one of any set of closely related people, since including closely related relatives can distort the results. And we remove outliers, people whose genetic ancestry doesn't seem to match up with their survey answers. To ensure a representative dataset, we filter aggressively — nearly ten percent of reference dataset candidates don't make the cut.

We also draw from public reference datasets, including the Human Genome Diversity Project, HapMap, and the 1000 Genomes project. Finally, we incorporate data from 23andMe-sponsored projects, typically collaborations with academic researchers. We perform the same filtering on these public and collaboration reference data that we do on the 23andMe customer data."

So theoretically, I could have been part of the control group before I took the test, as I considered all my four grandparents to be German (and they are, largely at least).

[Lucas]

@Harald Peterski, hello, brat! ;)

If you want to find Peterski just start thread about Germans from Poland:)

https://i.wpimg.pl/O/441x660/i.wp.pl/a/f/jpeg/29554/z_3997228.jpeg

[Lucas]

If you want to find Peterski just start thread about Germans from Poland:)

https://i.wpimg.pl/O/441x660/i.wp.pl/a/f/jpeg/29554/z_3997228.jpeg

As I expected:)

https://www.theapricity.com/forum/showthread.php?329944-People-from-Pomeranian-voivodeship-(mostly-Kashubians)-rank-classify-them&p=6844727&viewfull=1#post6844727