Scandinavian to British Isles or west/central Europe - New Algorithm

[hollifer]

There is a lengthy thread already about the change of origins with the new algorithm.

There appears to be a prevalent amount of people where the old algorithm showed a Scandinavia percentage, and the new one shows none or change to west/central Europe or increase in British Isles. I'm one of those people.

Before, I was close to 20% Scandinavian, 70+% eastern Europe (Polish), <10% BI, and 3% MidEast. Now, there's no Scandinavian at all, 73% eastern Europe, 10% BI, and 15% west/central Europe, with traces of Asia Minor and Iberia.

Close to half of my matches are Swedish. My paternal Grandma was Swedish descent from both her parents. My other 3 grandparents were Polish. So, my old results made more sense.

My main question is what is the historical explanation? I know that Vikings invaded much of Europe and mostly the BI. But, my recent ancestors going back at least to 3X great grandparents I know are Swedish. If there is any BI or west/central Europe ancestry, it goes further back and I have absolutely no idea, as I have no known recent ancestry from these areas.

[DaveInGreece]

There seems to be a major problem with the "Scandinavian" designation in the new version. Lots of people like you, with recent Scandinavian ancestry and a high Scandinavian percentage on the previous version (and or the calculators at Gedmatch etc.) now have little or no Scandinavian.

Conversely, people like me (purely British and Irish ancestry) are now showing high levels of Scandinavian - 22% for me! - which doesn't fit at all with known ancestry, the last version of My Origins, or any other ancestry estimator.

I've also suddenly jumped to having 29% "Iberian", whereas I had nothing vaguely southern European before.

It's normal for people with British ancestry to have a few per cent of Scandinavian (from Viking connections) and a lot of West/Central European (Anglo-Saxon and Norman), just as my sister does. But even a small amount of Iberian is rare and almost 30% is unheard of. The lack of anything S European showing for my full siblings (which they undoubtedly are) or for me on any test except this one makes this new My Origins seem totally unreliable.

[hollifer]

What is the best GEDmatch calculator to use to determine Scandinavian ancestry?

[dcx4610]

I went from 35% Scandinavian to 0% on FTDNA. Ancestry.com shows me at 15%.

All of the other calculators I've ran on gedmatch.com show a bit of Swedish DNA so I'd say it's probably closer to 15% than 35% but definitely not 0%.

I don't really understand the new calculator because I know others that are still showing Scandinavian DNA on 2.0 and it increased for some as well.

[Odis]

now we have to assume our past Population Finder-myOrigins-1 were all false..I went from 64% western/Euro to 0% lost all Scandinavian and shifted to basic all British 92% from 24%...do we now start to believe or wait 5 years for a results change.......

[dcx4610]

What is the best GEDmatch calculator to use to determine Scandinavian ancestry?

I like Dodecad V3 and Eurogenes K13. Dodecad feels more accurate to me but Eurogenes is useful to compare results. What you are looking for is Norwegian or Swedish results in the "Single Population Sharing" section. The closer the distance, the more DNA you have from that population.

[hollifer]

I like Dodecad V3 and Eurogenes K13. Dodecad feels more accurate to me but Eurogenes is useful to compare results. What you are looking for is Norwegian or Swedish results in the "Single Population Sharing" section. The closer the distance, the more DNA you have from that population.

Okay, so I just did these.

Dodecad V3:
Admix Results (sorted):

# Population Percent
1 West_European 44.16
2 East_European 25.04
3 Mediterranean 21.32
4 West_Asian 8.86
5 Northeast_Asian 0.45
6 East_African 0.18

Single Population Sharing:

# Population (source) Distance
1 Slovenian (Xing) 5.39
2 Hungarians (Behar) 7.62
3 German (Dodecad) 10.93
4 CEU (HapMap) 15.9
5 N._European (Xing) 16
6 Argyll (1000 Genomes) 16.08
7 Orcadian (HGDP) 16.99
8 Orkney (1000 Genomes) 17.19
9 FIN (1000Genomes) 18.9
10 Balkans (Dodecad) 20.07
11 Polish (Dodecad) 20.54
12 Mixed_Slav (Dodecad) 21.04
13 Mixed_Germanic (Dodecad) 22.71
14 French (Dodecad) 23.22
15 Romanians_14 (Behar) 23.46
16 French (HGDP) 23.55
17 Finnish (Dodecad) 23.67
18 Swedish (Dodecad) 23.92
19 Dutch (Dodecad) 24.36
20 Russian (Dodecad) 26.27

If I had no Swedish ancestry, can I assume that this wouldn't appear at all? These calculators' results are just based on the populations studied, is that correct? There are so many, that's it quite confusing.

Western European being higher than Eastern European is confusing, as I'm mostly Polish. But, I guess this is showing earlier DNA, from migration from the west perhaps.

Eurogenes K13:
Admix Results (sorted):

# Population Percent
1 Baltic 41.78
2 North_Atlantic 30.95
3 West_Med 13.45
4 West_Asian 6.65
5 East_Med 4.86
6 Amerindian 1
7 Northeast_African 0.76
8 Siberian 0.36
9 Oceanian 0.17
10 Sub-Saharan 0.03

Single Population Sharing:

# Population (source) Distance
1 South_Polish 3.04
2 Ukrainian 3.99
3 Ukrainian_Lviv 4.27
4 Polish 4.78
5 Croatian 7.5
6 Russian_Smolensk 8.11
7 Southwest_Russian 8.41
8 Ukrainian_Belgorod 8.73
9 Estonian_Polish 8.76
10 Belorussian 9.28
11 Hungarian 10.26
12 East_German 10.46
13 Moldavian 11.07
14 Southwest_Finnish 11.33
15 Estonian 11.79
16 Kargopol_Russian 12.12
17 Austrian 12.42
18 Lithuanian 12.93
19 Finnish 13.07
20 North_Swedish 14.9

This one seems more accurate, as Polish is closer. A grandfather was actually Ukrainian, from western Ukraine close to current Polish border, so I loop him into my Polish side.

I guess I can conclude from these that I do have Swedish DNA, just small percentages. That would make sense, as I just 1/4. Baltic is pretty high, so in the Scandinavian area. North Atlantic is BI, I presume. So, there is obviously Scandinavia and BI crossing over each other.

If you have any thoughts on my results above, I would love to hear it.

[hollifer]

I did the EUtest too. I think this one may be the most accurate. I think PL may be Polish. UA may be Ukraine. My understanding of how this works is 1 population measures the most ancestry I have. Since I'm mostly Polish, that matches. 4 populations measures each grandparent. But, there's no Polish there.

EUtest 4-Ancestors Oracle

Admix Results (sorted):

# Population Percent
1 SOUTH_BALTIC 24.65
2 NORTH-CENTRAL_EURO 24.34
3 EAST_EURO 16.94
4 ATLANTIC 16.65
5 WEST_MED 9.09
6 WEST_ASIAN 5.62
7 EAST_MED 2.51


Finished reading population data. 78 populations found.
13 components mode.

--------------------------------

Least-squares method.

Using 1 population approximation:
1 PL @ 9.580440
2 UA @ 11.235986
3 North_Swedish @ 11.339379
4 HU @ 11.800588
5 South_Finnish @ 12.435382
6 AT @ 12.435728
7 South_&_Central_Swedish @ 12.450314
8 West_Russian @ 12.496415
9 Ukrainian-Russian @ 13.415746
10 EE @ 13.671785
11 Belorussian @ 13.881100
12 NO @ 13.939371
13 West_&_Central_German @ 14.746617
14 NL @ 15.637305
15 DK @ 16.126081
16 East_Finnish @ 16.151129
17 Northwest_Russian @ 17.174822
18 East_Russian @ 17.535425
19 LIT @ 18.064743
20 English @ 18.207520

Using 2 populations approximation:
1 50% South_&_Central_Swedish +50% UA @ 3.584806


Using 3 populations approximation:
1 50% South_&_Central_Swedish +25% UA +25% UA @ 3.584806


Using 4 populations approximation:
++++++++++++++++++++++++++++++++++++
1 South_&_Central_Swedish + South_&_Central_Swedish + UA + UA @ 3.584806
2 NO + South_&_Central_Swedish + UA + UA @ 3.749105
3 South_&_Central_Swedish + UA + UA + West_&_Central_German @ 3.788461
4 NL + South_&_Central_Swedish + UA + UA @ 3.816091
5 LIT + South_&_Central_Swedish + UA + West_&_Central_German @ 3.867841
6 LIT + NO + UA + West_&_Central_German @ 3.889720
7 LIT + NL + South_&_Central_Swedish + UA @ 3.916969
8 LIT + NL + UA + West_&_Central_German @ 3.955134
9 NO + NO + UA + UA @ 3.961876
10 LIT + NL + NL + UA @ 3.968485
11 LIT + NL + NO + UA @ 3.973300
12 Belorussian + South_&_Central_Swedish + UA + West_&_Central_German @ 3.976090
13 HU + LIT + South_&_Central_Swedish + South_&_Central_Swedish @ 3.979797
14 PL + South_&_Central_Swedish + South_&_Central_Swedish + UA @ 4.008276
15 LIT + UA + West_&_Central_German + West_&_Central_German @ 4.014571
16 NO + UA + UA + West_&_Central_German @ 4.019593
17 Belorussian + NL + South_&_Central_Swedish + UA @ 4.030103
18 South_&_Central_Swedish + South_&_Central_Swedish + UA + Ukrainian-Russian @ 4.040820
19 English + LIT + South_&_Central_Swedish + UA @ 4.067041
20 NL + NO + UA + UA @ 4.078354

[dcx4610]

Population Percentage is basically the breakdown of DNA you have from each region.

The Single Population sharing is looking at your DNA as a whole and finding what population you match the closest to. The shorter the distance, the more native you are to that region.

No calculator is going to be perfect. For me, Dodecad lists my population as Argyll. My research leads me to Scotland so that makes sense. EUTest says I'm Dutch however and I can only find one Dutch grandmother.

It's best to use as many calculators as possible and try to match them to your research. There's going to be some confirmation bias involved I'm sure but if the research matches the results you are getting, it's at least some solid evidence.

[hollifer]

Well, the EUtest seems most accurate, as I'm mostly Polish (I assume PL is Polish), also Ukrainian (UA?), and Swedish.

I'm a redhead with many freckles. The freckles came from my mom's Polish/Ukrainian side. So, I'm figuring an ancestor from the BI migrated east over 300 years ago.

These tests are certainly fun to do, albeit there are many that creates confusion.

[khazaria]

Well, the EUtest seems most accurate, as I'm mostly Polish (I assume PL is Polish), also Ukrainian (UA?), and Swedish.

Yes, in EUtest I've confirmed that PL = Polish, UA = Ukrainian, LIT = Lithuanian, EE = Estonian, HU = Hungarian, RO = Romanian, AT = Austrian, IE = Irish

I'm a redhead with many freckles. The freckles came from my mom's Polish/Ukrainian side.

Which red hair SNP(s) are you positive for within your autosomal DNA?
Did you join Richard M. Stevens' Red Hair Variants Project at Family Tree DNA?

[hollifer]

Yes, in EUtest I've confirmed that PL = Polish, UA = Ukrainian, LIT = Lithuanian, EE = Estonian, HU = Hungarian, RO = Romanian, AT = Austrian, IE = Irish

Great, thanks for clarifying.

Which red hair SNP(s) are you positive for within your autosomal DNA?
Did you join Richard M. Stevens' Red Hair Variants Project at Family Tree DNA?

I have no idea. How do I check that? I'll look into that project.

[khazaria]

Which red hair SNP(s) are you positive for within your autosomal DNA?
Did you join Richard M. Stevens' Red Hair Variants Project at Family Tree DNA?

I have no idea. How do I check that? I'll look into that project.

The project is located at

https://www.familytreedna.com/groups/red-hair-variants/about

Below are some of the main values:

Arg160Trp = R160W = rs1805008
If you have one or two "T" letters for this value, you are eligible. That can be looked up inside your raw data file from Family Finder. When I opened my data field for this value, I saw I have "TT", which makes sense since I had reddish hair as a child. Kasia Bryc of 23andMe is a red-headed ethnic Pole who has this allele too.

R151C = rs1805007
If you score "T" in this, you are also eligible, but this SNP is not reported by Family Finder. You'd have to test for this on another site. Some ethnic Poles have at least one "T" there.

D294H = i3002507
If you score "C" in this, you are eligible, but this SNP isn't reported by Family Finder either.

Several others are listed in this screen:

https://www.familytreedna.com/public/RedHairVariants?iframe=yresults

[khazaria]

P.S. My dad's most accurate estimate in EUtest on line 13:
95.8% AJ
+ 4.2% PL
at genetic distance 3.78

is effectively identical to his estimate in MyOrigins 2.0:
95% Ashkenazi + <2% West Middle East
+ 4% East Europe

Did you know the EUtest calculator was created by a Pole?

[hollifer]

The project is located at

https://www.familytreedna.com/groups/red-hair-variants/about

Below are some of the main values:

Arg160Trp = R160W = rs1805008
If you have one or two "T" letters for this value, you are eligible. That can be looked up inside your raw data file from Family Finder. When I opened my data field for this value, I saw I have "TT", which makes sense since I had reddish hair as a child. Kasia Bryc of 23andMe is a red-headed ethnic Pole who has this allele too.

R151C = rs1805007
If you score "T" in this, you are also eligible, but this SNP is not reported by Family Finder. You'd have to test for this on another site. Some ethnic Poles have at least one "T" there.

D294H = i3002507
If you score "C" in this, you are eligible, but this SNP isn't reported by Family Finder either.

Several others are listed in this screen:
https://www.familytreedna.com/public...frame=yresults

Hmm. I checked on all RefSNPs on the project site and they're either CC or GG. I assure you I have red hair, auburn to be exact. I came out the womb with orange hair. What could this mean? My dad was dark brown hair, but his facial hair grew-in auburn, same exact color as me. He's half Swedish/half Polish. My mom had brownish-red hair and the freckles. She was half Polish/Ukrainian.