PDA

View Full Version : Eurogenes K13 & K15 - How far back are results consensus?



AnnieD
10-29-2016, 06:20 PM
Is there a consensus on how far back the results for Gedmatch Eurogenes calculators K13 and K15 go? I've seen comments on autosomal discussion forums that Eurogenes results are older than commercial DNA tests such as 23andMe (~ 500 years?) & FTDNA (~1k years).

Opinions based on your results & known family tree or any technical knowledge are welcome. I will then tally our "consensus".

kikkk
10-29-2016, 07:14 PM
Yes both k13&k15, unlike DNA LAND&FTDNA, detected some minor South Asian (on my autosomal makeup) that was confirmed, later on, by my Dr McDonald chromosomes paintings.
For the oracles k13 is closer than k15 , as my European side seems to be Swedish-Irish mix.

K13
# Population Percent
1 North_Atlantic 32.12
2 West_Med 15.81
3 East_Med 13.86
4 Baltic 13.59
5 Red_Sea 10.00
6 Northeast_African 4.98
7 West_Asian 4.24
8 Sub-Saharan 3.47

Population
North_Atlantic 32.12
Baltic 13.59
West_Med 15.81
West_Asian 4.24
East_Med 13.86
Red_Sea 10.00
South_Asian 0.49
East_Asian 0.51
Siberian -
Amerindian 0.66
Oceanian 0.28
Northeast_African 4.98
Sub-Saharan 3.47

Finished reading population data. 204 populations found.
13 components mode.

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

Least-squares method.

Using 1 population approximation:
1 Portuguese @ 14.018429
2 Spanish_Galicia @ 14.506813
3 Spanish_Extremadura @ 15.099699
4 North_Italian @ 15.172968
5 French @ 15.460756
6 Spanish_Castilla_Y_Leon @ 15.869692
7 Spanish_Murcia @ 16.253592
8 Spanish_Cataluna @ 17.002790
9 Tuscan @ 17.456627
10 Spanish_Valencia @ 18.253286
11 Serbian @ 18.322350
12 Spanish_Castilla_La_Mancha @ 18.377823
13 Spanish_Andalucia @ 18.469498
14 Romanian @ 18.583807
15 West_German @ 18.740555
16 Spanish_Cantabria @ 19.541889
17 South_Dutch @ 19.605225
18 Bulgarian @ 20.469122
19 Austrian @ 21.005989
20 Southwest_French @ 21.494698

Using 2 populations approximation:
1 50% Algerian +50% Norwegian @ 3.514825


Using 3 populations approximation:
1 50% Algerian +25% Irish +25% Norwegian @ 3.264417


Using 4 populations approximation:
1 Algerian + Irish + Moroccan + Norwegian @ 3.188251
2 Algerian + Moroccan + Norwegian + West_Scottish @ 3.210564
3 Algerian + Algerian + Irish + Norwegian @ 3.264417
4 Algerian + Moroccan + Norwegian + Norwegian @ 3.288047
5 Algerian + Algerian + Norwegian + West_Scottish @ 3.336408
6 Algerian + Moroccan + North_Dutch + Norwegian @ 3.346681
7 Algerian + Moroccan + Swedish + West_Scottish @ 3.446958
8 Moroccan + Norwegian + Tunisian + West_Scottish @ 3.464065
9 Irish + Moroccan + Norwegian + Tunisian @ 3.470368
10 Algerian + Danish + Moroccan + Norwegian @ 3.483180
11 Algerian + Irish + Moroccan + Swedish @ 3.498240
12 Algerian + Irish + Moroccan + North_Dutch @ 3.498912
13 Algerian + Algerian + Norwegian + Norwegian @ 3.514825
14 Algerian + Algerian + Swedish + West_Scottish @ 3.532673
15 Algerian + Moroccan + North_Dutch + West_Scottish @ 3.533063
16 Algerian + Algerian + Irish + North_Dutch @ 3.540893
17 Moroccan + Moroccan + Norwegian + Norwegian @ 3.542603
18 Algerian + Algerian + Irish + Swedish @ 3.544111
19 Algerian + Algerian + North_Dutch + Norwegian @ 3.544520
20 Algerian + Moroccan + Norwegian + Orcadian @ 3.568996

K15
1 North_Sea 25.44
2 Atlantic 17.00
3 West_Med 12.50
4 East_Med 11.39
5 Red_Sea 9.98
6 Baltic 6.28
7 Northeast_African 5.48
8 West_Asian 3.69
9 Eastern_Euro 3.69
10 Sub-Saharan 3.30

Population
North_Sea 25.44
Atlantic 17.00
Baltic 6.28
Eastern_Euro 3.69
West_Med 12.50
West_Asian 3.69
East_Med 11.39
Red_Sea 9.98
South_Asian 0.31
Southeast_Asian 0.48
Siberian -
Amerindian 0.39
Oceanian 0.06
Northeast_African 5.48
Sub-Saharan 3.30
Finished reading population data. 207 populations found.
15 components mode.

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

Least-squares method.

Using 1 population approximation:
1 Spanish_Galicia @ 14.588771
2 French @ 14.938599
3 Portuguese @ 15.891663
4 West_German @ 16.730318
5 Spanish_Extremadura @ 17.254105
6 North_Italian @ 17.709951
7 Serbian @ 17.949858
8 Spanish_Castilla_Y_Leon @ 18.032417
9 South_Dutch @ 18.122677
10 Spanish_Murcia @ 18.458818
11 Spanish_Cataluna @ 18.485561
12 Tuscan @ 19.544058
13 Romanian @ 19.794596
14 Hungarian @ 20.520821
15 Austrian @ 21.227255
16 East_German @ 21.338644
17 Greek_Thessaly @ 21.864101
18 Bulgarian @ 22.011896
19 Southwest_English @ 22.133476
20 Spanish_Cantabria @ 22.233433

Using 2 populations approximation:
1 50% Algerian +50% West_Norwegian @ 4.978865


Using 3 populations approximation:
1 50% Algerian +25% West_Norwegian +25% West_Norwegian @ 4.978865


Using 4 populations approximation:
1 Algerian + Algerian + West_Norwegian + West_Norwegian @ 4.978865
2 Algerian + Moroccan + West_Norwegian + West_Norwegian @ 5.071027
3 Algerian + Algerian + Orcadian + West_Norwegian @ 5.157695
4 Algerian + Algerian + North_Dutch + West_Norwegian @ 5.387486
5 Algerian + Moroccan + Orcadian + West_Norwegian @ 5.403839
6 Algerian + Tunisian + West_Norwegian + West_Norwegian @ 5.427441
7 Algerian + Algerian + Norwegian + West_Norwegian @ 5.455012
8 Moroccan + Tunisian + West_Norwegian + West_Norwegian @ 5.456275
9 Algerian + Mozabite_Berber + West_Norwegian + West_Norwegian @ 5.506931
10 Algerian + Algerian + Swedish + West_Norwegian @ 5.521755
11 Moroccan + Moroccan + West_Norwegian + West_Norwegian @ 5.529415
12 Algerian + Algerian + West_Norwegian + West_Scottish @ 5.531826
13 Algerian + Algerian + Orcadian + Swedish @ 5.534675
14 Ashkenazi + Moroccan + Orcadian + West_Norwegian @ 5.560763
15 Algerian + Moroccan + Norwegian + West_Norwegian @ 5.588689
16 Algerian + Moroccan + Swedish + West_Norwegian @ 5.601461
17 Algerian + Moroccan + North_Dutch + West_Norwegian @ 5.606465
18 Algerian + Algerian + Norwegian + Orcadian @ 5.608943
19 Algerian + Algerian + Irish + West_Norwegian @ 5.643203
20 Ashkenazi + Moroccan + West_Norwegian + West_Norwegian @ 5.668918

AnnieD
10-29-2016, 08:57 PM
Yes both k13&k15, unlike DNA LAND&FTDNA, detected some minor South Asian (on my autosomal makeup) that was confirmed, later on, by my Dr McDonald chromosomes paintings.
For the oracles k13 is closer than k15 , as my European side seems to be Swedish-Irish mix.

You are 1 of the lucky ones to get a chromosome painting by Dr. McDonald before he stopped offering to the masses. By the time I came along to genetic genealogy, I also missed out on Davidski's (Eurogenes) Personal Genetic Detective Services. However, I did catch his Eurogenes K8 deeper ancestry calculator which he offered via $20 contribution (not on Gedmatch).

FredBats
10-30-2016, 10:39 AM
It'd be interesting on the distance. But how accurate is the reference populations?

I have a few European friends and it certainly raised an eyebrow when the Russian and the Finn shared something in common - overwhelming Lithuanian with just the fst distance separating the Russian [closer] and the Finn [far more distant].


As it is, I don't see anything outstandingly distant. Dad as his ancestry is well researched both sides.

k13
Using 4 populations approximation:
1 Norwegian + Southeast_English + Spanish_Cataluna + West_German @ 1.607540
2 North_Dutch + Norwegian + French_Basque + West_German @ 1.687943
3 Danish + Norwegian + Spanish_Aragon + West_German @ 1.709811

Norwegian & Danish - obvious, Northern Brit
Spanish_Cataluna & Aragon - southern French
West_German - French border ancestry maybe older but doubtful


k15
1 Irish + Spanish_Cantabria + Swedish + West_German @ 1.947187
2 Irish + Spanish_Aragon + Swedish + West_German @ 2.062594
3 Spanish_Valencia + Swedish + West_German + West_Scottish @ 2.093920
4 Irish + North_Dutch + Spanish_Cataluna + West_German @ 2.097420

Irish - irish ancestry
Swedish - Brit
North_Dutch - Brit possibly French influx
Spanish_Cataluna & Aragon - southern French
West_German - maybe Brit / but likely French influx

The lack of northern Italian [and as per the population search north Italian is included] was amusing as northern Italian is old ancestry on his mother's side. Shows up on other sites.

It was even more amusing when his son, my brother, is apparently overwhelmingly West_Scottish with no German despite how mum's ancestry is German/Irish as well. :rolleyes:


Interestingly both Geno 2.0 and FTDNA show what I'd say is older ancestry than what k13 & k15 shows. The only thing k13 & k15 can be considered old is this degree is fact that he does register as eastern European [3.43] on k15 alone.

Regardless as a comparable Geno 2.0 puts him as primary Danish & secondary Greek [with the percentages almost neck & neck]. FTDNA has hits in the origins [not the matches] for his mtDNA as been Moldavian and obviously the Moldavian region does factor in because Greece is just a little west of Moldova & Moldova itself considered eastern European.

However, though k15 indicates eastern european it actually doesn't identify any. Not in oracle 4 or mixed. In that sense other oracles who have in passing mentioned Russian, Bulgarian and Ukraine could be considered more accurate to actual distant ancestry.



The irony is most people say phased kits are useless well I found the phased kit for dad far better here. What with Portuguese & La_Brana-1 been some of those included. Spanish regions that aren't bordering Basque country or the French border. The inclusion of Afghan in the phased kits is interesting. And the presence of Moldavian made me chuckle because as mentioned above his maternal DNA on FTDNA shows a few Moldavian "hits" on the origins not matches.

AnnieD
10-30-2016, 03:52 PM
I had a go at the Eurogenes blog yesterday to find any references to how far back the K13 or K15 measured to no avail. :( I will try again today as the blog is rather extensive.

As a fun exercise of the possible vagaries of comparing different tools, my father, myself & a fully European British member on this forum all wind up as "SE English" on Eurogenes K13 single pop mode. However, all 3 of our FTDNA (1k years back?) results are strikingly different, especially in Scandinavian & S. Europe categories. My father's results are only at FTDNA, but when I compare myself to the British tester with 23andMe (500 years back?), strikingly different again. :confused:

Therefore, I'm curious what interpretative process testers use when they compare these commercial tests to an admixture calculator like Eurogenes K13 or K15? It seems that we would need to have an idea of how far back each tool purports to measure. I will ask this question on Eurogenes threads & see if any takers. ;)

firemonkey
10-30-2016, 04:21 PM
I(English,Scottish and Irish) wind up with West Scottish(Ancestry), West Scottish(FTDNA) and West Scottish(23andMe) with K13 single pop sharing.
My father (English, FTDNA) winds up with Orcadian.

AnnieD
10-30-2016, 04:36 PM
It'd be interesting on the distance. But how accurate is the reference populations?

I have a few European friends and it certainly raised an eyebrow when the Russian and the Finn shared something in common - overwhelming Lithuanian with just the fst distance separating the Russian [closer] and the Finn [far more distant].

I have seen some of the Eurogenes & experimental calculator results of a nearly fully Finnish poster & they lived up to, at least in my mind, the "genetic outlier" that I've seen in AC discussions. I blame this "hobby" today on my DNA Tribes 26-marker STR test which placed my "deep" ancestry as Finnish and primary regional category as Baltic / Slavic. I didn't know my family tree well enough to tell any test that it was "wrong". However, it made me wonder how "deep" the test measured ancestry. :noidea:

Fast forward a year & I took the same company's "deep" SNP analysis ... back in NW Europe with Cornish, English & Argyll results. From "deep" NE Euro to NW Euro ... it happens if you subscribe to current population genetic theories. :D However, as a non-scientist consumer, how do I interpret the generic "deep" results across different populations without any quantifying data?

If interested in seeing how a commercial test defends its methodology, I've copied their customer service response below:

"The differences between our tests can be summarized as follows: 1) they are based on different markers that can both reflect fairly deep ancestry (dating back hundreds and potentially thousands of years, we cannot qualify more precisely); 2) the reference data used is not the same: STR population database is much more complete with respect to geographic coverage but the SNP analysis tends to be more precise at the cost of smaller reference datasets. Thus, while generally customers get fairly consistent results using STR and SNP tests, once in a while they do get different results as the tests may reflect different "portions" of ancestry which, in combination, may provide a fuller picture."

FredBats
10-30-2016, 10:02 PM
I have seen some of the Eurogenes & experimental calculator results of a nearly fully Finnish poster & they lived up to, at least in my mind, the "genetic outlier" that I've seen in AC discussions.

Oh he's outlandish.

Finn
1 Lithuanian @ 26.165739
2 Erzya @ 30.073460
3 Estonian @ 30.311512

1 Lithuanian + Lithuanian + Lithuanian + Lithuanian @ 26.165739
2 Erzya + Lithuanian + Lithuanian + Lithuanian @ 26.747993
3 Estonian + Lithuanian + Lithuanian + Lithuanian @ 27.126207


Russian
1 Lithuanian @ 5.579403
2 Estonian @ 9.296667
3 Belorussian @ 10.815320

1 Lithuanian + Lithuanian + Lithuanian + Lithuanian @ 5.579403
2 Estonian + Lithuanian + Lithuanian + Lithuanian @ 6.133565
3 East_Finnish + Lithuanian + Lithuanian + Lithuanian @ 6.665838



But this is what I mean by the accuracy. This is Eurogenes K13 for my "sink your teeth into" post in the autosomal. He is very English don't you think?

Using 4 populations approximation:
1 West_Scottish + West_Scottish + West_Scottish + West_Scottish @ 22.576336
2 Irish + West_Scottish + West_Scottish + West_Scottish @ 22.778667
3 Southwest_English + West_Scottish + West_Scottish + West_Scottish @ 22.801132
4 French_Basque + West_Scottish + West_Scottish + West_Scottish @ 22.926781
5 Irish + Irish + West_Scottish + West_Scottish @ 22.984077

AnnieD
10-30-2016, 10:31 PM
Oh he's outlandish.

Finn
1 Lithuanian @ 26.165739
2 Erzya @ 30.073460
3 Estonian @ 30.311512

1 Lithuanian + Lithuanian + Lithuanian + Lithuanian @ 26.165739
2 Erzya + Lithuanian + Lithuanian + Lithuanian @ 26.747993
3 Estonian + Lithuanian + Lithuanian + Lithuanian @ 27.126207


Russian
1 Lithuanian @ 5.579403
2 Estonian @ 9.296667
3 Belorussian @ 10.815320

1 Lithuanian + Lithuanian + Lithuanian + Lithuanian @ 5.579403
2 Estonian + Lithuanian + Lithuanian + Lithuanian @ 6.133565
3 East_Finnish + Lithuanian + Lithuanian + Lithuanian @ 6.665838


But this is what I mean by the accuracy. This is Eurogenes K13 for my "sink your teeth into" post in the autosomal. He is very English don't you think?

Using 4 populations approximation:
1 West_Scottish + West_Scottish + West_Scottish + West_Scottish @ 22.576336
2 Irish + West_Scottish + West_Scottish + West_Scottish @ 22.778667
3 Southwest_English + West_Scottish + West_Scottish + West_Scottish @ 22.801132
4 French_Basque + West_Scottish + West_Scottish + West_Scottish @ 22.926781
5 Irish + Irish + West_Scottish + West_Scottish @ 22.984077

Wow, those are gigantic FST for the Finn & British testers. I've read that the ideal threshold is < 2.0. At least they landed in reasonably approximate geographic regions. ;)

This fully W. Scottish result @ 4-pop mode but at such a high FST is a first for British results that I've seen:

Using 4 populations approximation:
1 West_Scottish + West_Scottish + West_Scottish + West_Scottish @ 22.576336