PDA

View Full Version : Best or most accurate Gedmatch Admixture?



Jutlander
12-30-2014, 09:11 PM
I am new to this forum and the concept of DNA Ancestry, i have only turned to it recently when my paper trails hit dead ends.

Please can anything suggest which out of Eurogenes, MDLP and Dodecad is the most accurate or reliable based on results that make sense to members own Ethnic Ancestry?

Also whether the Oracle or Oracle-4 is better than the other, i seem to gain crazy results that do not add up on the Gedmatch basic oracle but better hits on the Aleksandr Burnashev version.

All opinions professional or otherwise would be of great help.

dp
01-06-2015, 07:40 PM
Here is a comparative example, drawn from Oracle2 runs. Basically I think admixture runs, for a person with ancestors from the same geographical region, are accurate to only about 500 miles in radius from their predicted coordinate. Of course if your ancestry has people from different ethnicities the results will be skewed. That's when Oracle4 becomes useful.
http://www.anthrogenica.com/attachment.php?attachmentid=2290&d=1407531220
Hope this gives you an idea. Check out the mapping thread for the following post (http://www.anthrogenica.com/showthread.php?1778-Mapping-Your-Location-(GEDmatch-calculators)&p=48134&viewfull=1#post48134)
dp :-)
PS: Welcome to Anthrogenica.

vettor
01-07-2015, 02:54 AM
I would like to see a Gedmatch mixture which corrects/adjusts itself as time goes by instead of seeing brand new admixture parameters all the time.

Tolan
01-07-2015, 09:27 AM
There are individual variability. You may even have slightly different results with your siblings.

All my ancestors come from the same (small) geographical area, and all the calculators give me this area 800 km (500 miles) further east than reality.
The problem for results from Oracle, it is the reference populations.
I discovered, for example, that the reference to French DODECAD are all from eastern France (Picardy to RhoneAlpes), andr even two Belgians, one Luxembourger, one Swiss and one Italian, one german in the French sample!
And also people with unknown fathers...
This is not representative of a French "average"!
Rather than having new calculators, I would rather have more accurate and precise reference populations!

rms2
01-07-2015, 05:24 PM
Here is a comparative example, drawn from Oracle2 runs. Basically I think admixture runs, for a person with ancestors from the same geographical region, are accurate to only about 500 miles in radius from their predicted coordinate. Of course if your ancestry has people from different ethnicities the results will be skewed. That's when Oracle4 becomes useful.
http://www.anthrogenica.com/attachment.php?attachmentid=2290&d=1407531220
Hope this gives you an idea. Check out the mapping thread for the following post (http://www.anthrogenica.com/showthread.php?1778-Mapping-Your-Location-(GEDmatch-calculators)&p=48134&viewfull=1#post48134)
dp :-)
PS: Welcome to Anthrogenica.

I remember when we did those maps. Yours looks very good. Mine was sloppier, but I think it yielded a fairly accurate result, given what I know of the totality of my ancestry.

3335

So, I think your recommendation was excellent.

randwulf
01-07-2015, 11:34 PM
The map idea is pretty interesting. I ran several calculators and drew the points by hand. It created a triangle from about Birmingham, UK to Berlin to North Italy. The resulting triangle pretty well encompasses my known ancestry with just a couple slightly outside the triangle, but not by much. The center is about where I would expect it to be, as well (Eastern France to West/South Germany).

Solothurn
01-24-2015, 07:08 PM
I think I should do this for my wife as FTDNA's My Origins gives:

68% Scandinavian
23% West and Central Europe
9% British

Actually my wife is 100% 'known' British ancestry.

Any thoughts???

randwulf
01-24-2015, 07:45 PM
I think I should do this for my wife as FTDNA's My Origins gives:

68% Scandinavian
23% West and Central Europe
9% British

Actually my wife is 100% 'known' British ancestry.

Any thoughts???

It worked really well for mine, as stated above. I did it by hand on a paper map and just took several of the calculators two population results. A definite shape/area came from it that was really accurate including an almost perfect "center", even though the individual lines weren't always specific to my known ancestry.

DMXX
01-24-2015, 07:55 PM
At present, for West Eurasians, I'd consider the MDLP K23b to be the most informative calculator for two reasons:

1) Superior reference population database. Many of the calculators at GEDmatch are several years old and don't feature newer ones.
2) Highly accurate Oracle results, as seen here (http://www.anthrogenica.com/showthread.php?3159-Post-your-MDLP-K23b-calculator-Oracle-ancestry-here).

Vadim personally recalibrated the calculator several times following user feedback before it was uploaded to GEDmatch in that thread, which was both a highly unprecedented and positive move.

CelticGerman
01-24-2015, 08:41 PM
At present, for West Eurasians, I'd consider the MDLP K23b to be the most informative calculator for two reasons:

1) Superior reference population database. Many of the calculators at GEDmatch are several years old and don't feature newer ones.
2) Highly accurate Oracle results, as seen here (http://www.anthrogenica.com/showthread.php?3159-Post-your-MDLP-K23b-calculator-Oracle-ancestry-here).

Vadim personally recalibrated the calculator several times following user feedback before it was uploaded to GEDmatch in that thread, which was both a highly unprecedented and positive move.

Most convincing calculators are EUROGENES K13 and K15 in my opinion. At least results are in line with my known ancestry. DODECAD K12b is quite good as well.

MDLP K23b creates some headache: On the one hand there are interesting new samples from Frisia and Belgium, but on the other hand there are still South Germans close to top in my case, which is different from other calculators and completely wrong compared to my known ancestry. Austria is too close as well. Similar doubts when looking at results of my wife who got French on top with every calculator, which is in line with her ancestry. However, K23b results in North Italian, German Volga, South German and only at rank 4 French with distance 10.7.

CelticGerman
01-24-2015, 09:03 PM
I think I should do this for my wife as FTDNA's My Origins gives:

68% Scandinavian
23% West and Central Europe
9% British

Actually my wife is 100% 'known' British ancestry.

Any thoughts???

Here are my thoughts: FTDNA is simply not able to make the difference between West and Central Europe, Scandinavian and British correctly (it is difficult, I agree). I'm German and I got 58% Scandinavian, 18% Eastern Europe, 13% British Isles and 11% Southern Europe. No West and Central Europe at all, region where 97% of my known ancestry lived. I think almost all my North German ancestry is considered as Scandinavian here. The comparison of results of my four children is very interesting as well: Scandinavia varies from 18% to 50%; Eastern Europe from 0 to 23%; Southern Europe from 5 to 36%; British Isles from 9 to 29%; Western and Central Europe three children simply 0 and one child 45%! A lot of work to do FTDNA. 23andMe did more progress in my opinion.

randwulf
01-25-2015, 10:21 PM
To answer the original thread question, for me being of American mix (if I had to say it with big crayons 70% German, 30% English), almost all of them picked for a single population French with one German pick and one SE English pick. These were all pretty good guesses or ways to deal with the mix, I think. The winner at one population was Eurogenes EUTest V2 K15 which picked French or West German at close to the same "distance". At two populations, the best was MDLP K23b which picked me as 50% French and 50% South German. Most of the others picked me as a mix of a West Mediterranean with a Scandinavian for two populations. None of them were particularly good at three or four populations, most thinking I was a Spanish/East European/Scandinavian/Italian to Greek Mediterranean mix of some form.

Geno 2.0 picked me as British or German with a 41% Mediterranean, 40% Northern European 17% SW Asian, and 2% NE Asian mix. FTDNA gave me 42% Scandinavian, 38% Southern Europe, and 20% West/Central Europe. I suppose these are reasonable. Geno 2.0 and FTDNA are coloring with really big crayons.

Jutlander
01-26-2015, 02:15 PM
I think after using what i have read to be the top admixtures i think each one just distorts each others. My Oracle 4 charts are all over the place but some themes do reoccur now and again.

Eurogenes EU-Test

1 DK + DK + DK + French_Basque @ 2.668931
2 DK + DK + French_Basque + IE @ 2.891795
3 DK + DK + French_Basque + Orcadian @ 2.893455
4 DK + DK + French_Basque + NO @ 2.912057
5 DK + DK + French_Basque + Scottish @ 2.963554
6 DK + French_Basque + IE + NO @ 2.995116
7 DK + French_Basque + NO + Orcadian @ 3.039167
8 DK + French_Basque + NO + Scottish @ 3.057212
9 DK + DK + English + French_Basque @ 3.101458
10 DK + English + French_Basque + NO @ 3.107097
11 DK + French_Basque + IE + South_&_Central_Swedish @ 3.214302
12 DK + DK + French_Basque + South_&_Central_Swedish @ 3.236929
13 DK + French_Basque + Scottish + South_&_Central_Swedish @ 3.277608
14 DK + French_Basque + Orcadian + South_&_Central_Swedish @ 3.294412
15 Cornish + Cornish + Cornish + IE @ 3.315875
16 DK + DK + French_Basque + NL @ 3.323825
17 English + French_Basque + NO + NO @ 3.361999
18 DK + French_Basque + IE + North_Swedish @ 3.368464
19 French_Basque + IE + NO + NO @ 3.369031
20 DK + English + French_Basque + South_&_Central_Swedish @ 3.374438

EU- Test V2

1 French_Basque + West_Scottish + West_Scottish + West_Scottish @ 4.565075
2 French_Basque + Irish + West_Scottish + West_Scottish @ 4.613229
3 Danish + French_Basque + West_Scottish + West_Scottish @ 4.657774
4 French_Basque + Irish + Irish + West_Scottish @ 4.692403
5 Danish + French_Basque + Irish + West_Scottish @ 4.692967
6 French_Basque + North_German + West_Scottish + West_Scottish @ 4.737082
7 Danish + French_Basque + Irish + Irish @ 4.767036
8 French_Basque + Irish + Irish + Irish @ 4.801063
9 French_Basque + Irish + North_German + West_Scottish @ 4.847809
10 French_Basque + Orcadian + West_Scottish + West_Scottish @ 4.914096
11 French_Basque + Irish + Orcadian + West_Scottish @ 4.919183
12 French_Basque + Southeast_English + West_Scottish + West_Scottish @ 4.923421
13 Danish + Danish + French_Basque + West_Scottish @ 4.930357
14 French_Basque + North_German + Orcadian + West_Scottish @ 4.933901
15 French_Basque + Irish + Irish + Orcadian @ 4.957059
16 Danish + Danish + French_Basque + Irish @ 4.964528
17 Danish + French_Basque + Orcadian + West_Scottish @ 4.976859
18 Danish + French_Basque + Irish + Orcadian @ 4.978167
19 French_Basque + Irish + Irish + North_German @ 4.992580
20 French_Basque + Irish + Southeast_English + West_Scottish @ 4.998106


Dodecad V3

1 Balkans + Cornwall + Kent + Kent @ 1.531679
2 French + Argyll + Argyll + Argyll @ 1.543706
3 Balkans + British + Kent + Kent @ 1.556876
4 Balkans + British_Isles + Kent + Kent @ 1.623801
5 Balkans + British + British + Kent @ 1.666041
6 Balkans + Kent + Kent + Kent @ 1.672940
7 Balkans + British + Cornwall + Kent @ 1.689611
8 Balkans + British_Isles + Cornwall + Kent @ 1.697067
9 Balkans + British + British_Isles + Kent @ 1.713086
10 French + Argyll + Argyll + N._European @ 1.730409
11 Balkans + British + Cornwall + Dutch @ 1.732935
12 British_Isles + French + Mixed_Germanic + Slovenian @ 1.733330
13 Balkans + Cornwall + Cornwall + Dutch @ 1.741696
14 Balkans + Cornwall + Cornwall + Kent @ 1.750032
15 Balkans + British_Isles + Cornwall + Dutch @ 1.755064
16 Balkans + British + British + Dutch @ 1.759293
17 Balkans + Cornwall + Dutch + Kent @ 1.766985
18 Balkans + British_Isles + British_Isles + Kent @ 1.774372
19 French + German + Mixed_Germanic + Argyll @ 1.788904
20 Balkans + British + British_Isles + Dutch @ 1.798341

MDLP K23b

1 French + Frisian + Frisian + Frisian @ 1.553141
2 French + Frisian + Frisian + Irish @ 1.627798
3 Belgian + French + Frisian + Frisian @ 1.644677
4 Belgian + French + Frisian + Irish @ 1.669494
5 French + Frisian + Frisian + North_European @ 1.674586
6 Dutch + English_Kent_GBR + French + Frisian @ 1.685138
7 English_Cornwall_GBR + English_Kent_GBR + Frisian + Frisian @ 1.700366
8 Dutch + French + Frisian + Frisian @ 1.705611
9 British + Dutch + French + Frisian @ 1.714180
10 English + French + Frisian + Frisian @ 1.719059
11 English_Kent_GBR + English_Kent_GBR + Frisian + Irish @ 1.732669
12 Dutch + French + Frisian + Irish @ 1.775222
13 Dutch + English_Cornwall_GBR + French + Frisian @ 1.775391
14 English_Kent_GBR + English_Kent_GBR + Frisian + Frisian @ 1.780417
15 Belgian + English_Cornwall_GBR + English_Kent_GBR + Frisian @ 1.782306
16 CEU + Dutch + French + Frisian @ 1.782576
17 English_Kent_GBR + Frisian + Frisian + Welsh @ 1.788402
18 English_Cornwall_GBR + English_Cornwall_GBR + Frisian + Frisian @ 1.791348
19 Belgian + English_Cornwall_GBR + English_Cornwall_GBR + Frisian @ 1.796403

I gather the rules regarding what population references are submitted to cross reference with the tests and that it only shows the best fit, its just that each one is changing too much to the point that i can't mean much. For instance the Eu Test and the V2 changed everything of Norwegian and Swedish and came up with Orcadian and West Scottish? Anyone else understand all this?

Tolan
01-27-2015, 09:03 AM
With Oracle at 3 or 4 populations, there are many more possibilities, and therefore will be less relevant!
The genetic difference between the proposals data will not be huge, and the first line will not be very far from the 10th

Jutlander
01-27-2015, 02:38 PM
Dodecad and MDLP seem to have an obsession about throwing in Cornwall, but as yet i have no known link on paper or theoretical to Cornish whatsoever.

Agamemnon
01-27-2015, 03:00 PM
Here's what my mother gets on the 4-way oracles of several tests:

MDLP K23b 4-Ancestors Oracle:

1 British + Croat_BH + English_Cornwall_GBR + North_European @ 0.783550
2 Dutch + Dutch + Kashub + Spanish_Cataluna_IBS @ 0.785721
3 British + English + Norwegian_East + Serb_Serbia @ 0.787961
4 English_Cornwall_GBR + North_European + Norwegian_East + Serb_Serbia @ 0.822201
5 Belgian + Belgian + Dutch + South_German @ 0.840481
6 North_European + Norwegian_East + Serb_Serbia + Welsh @ 0.855515
7 English + English_Cornwall_GBR + Frisian + Slovenian @ 0.856773
8 British + Croat_BH + English + Welsh @ 0.863890
9 English + English_Cornwall_GBR + Frisian + Hungarian @ 0.878623
10 Croat_BH + English_Cornwall_GBR + English_Cornwall_GBR + North_European @ 0.879899
11 Belgian + Dutch + Frisian + South_German @ 0.889204
12 Belgian + English + English_Cornwall_GBR + Slovenian @ 0.900993
13 English + Norwegian_East + Serb_Serbia + Welsh @ 0.911044
14 Croat_BH + English_Cornwall_GBR + North_European + Welsh @ 0.914970
15 CEU + English_Kent_GBR + Norwegian_East + Serb_Serbia @ 0.921902
16 British + North_European + Norwegian_East + Serb_Serbia @ 0.923345
17 Belarusian-East + Dutch + Frisian + Spanish_Cataluna_IBS @ 0.924610
18 Bosnian + British + British + North_European @ 0.924696
19 CEU + Croat_BH + English_Cornwall_GBR + North_European @ 0.926289
20 Frisian + German-Volga + Norwegian_East + South_German @ 0.930812


Dodecad V3 4-Ancestors Oracle:

1 German + Spaniards + Swedish + Argyll @ 1.023146
2 French_Basque + Mixed_Germanic + Swedish + Slovenian @ 1.113911
3 German + IBS + Swedish + Argyll @ 1.210174
4 German + Spanish + Swedish + Argyll @ 1.238664
5 German + Spaniards + Swedish + N._European @ 1.244013
6 German + Orcadian + Portuguese + Swedish @ 1.258096
7 German + Orkney + Portuguese + Swedish @ 1.295823
8 French_Basque + Hungarians + Mixed_Germanic + Swedish @ 1.299515
9 German + Orkney + Spaniards + Swedish @ 1.329143
10 French + German + Orkney + Orkney @ 1.331661
11 German + German + Norwegian + Spaniards @ 1.341599
12 Hungarians + Irish + Spaniards + Swedish @ 1.351833
13 German + Portuguese + Swedish + Argyll @ 1.351872
14 IBS + Norwegian + Norwegian + Slovenian @ 1.352066
15 German + Orcadian + Spaniards + Swedish @ 1.356080
16 French_Basque + Hungarians + Mixed_Germanic + Norwegian @ 1.356163
17 French + German + Orcadian + Orkney @ 1.379786
18 German + IBS + Orkney + Swedish @ 1.403893
19 French_Basque + Norwegian + N._European + Slovenian @ 1.410785
20 French_Basque + Norwegian + Argyll + Slovenian @ 1.416492


Dodecad K7b 4-Ancestors Oracle:

1 British + British + English + French @ 0.000000
2 British + British + English + French @ 0.000000
3 British + British_Isles + English + French @ 0.000000
4 British + Cornwall + French + Orkney @ 0.000000
5 British + English + French + Irish @ 0.000000
6 British + French + Irish + Kent @ 0.000000
7 British + French + Kent + Orcadian @ 0.000000
8 British_Isles + British_Isles + English + French @ 0.000000
9 British_Isles + British_Isles + French + Kent @ 0.000000
10 British_Isles + English + French + Irish @ 0.000000
11 British_Isles + French + Irish + Kent @ 0.000000
12 English + English + French + Orkney @ 0.000000
13 English + French + Kent + Orkney @ 0.000000
14 English + French + Irish + Irish @ 0.252067
15 British_Isles + Cornwall + French + Orkney @ 0.252786
16 British + English + French + Orcadian @ 0.256227
17 British_Isles + British_Isles + French + Kent @ 0.256377
18 British + British_Isles + English + French @ 0.256502
19 British + English + French + Irish @ 0.259544
20 British_Isles + English + French + Irish @ 0.259544


Dodecad K12b 4-Ancestors Oracle:

1 Aragon + Belorussian + CEU30 + Orkney @ 0.000000
2 Aragon + Belorussian + English + Orkney @ 0.000000
3 Argyll + Belorussian + British_Isles + Valencia @ 0.000000
4 Belorussian + British + British + Cataluna @ 0.000000
5 Belorussian + British_Isles + Irish + Valencia @ 0.000000
6 Belorussian + British_Isles + Orcadian + Valencia @ 0.000000
7 Belorussian + Cantabria + CEU30 + Irish @ 0.000000
8 Belorussian + Cantabria + English + Irish @ 0.000000
9 Belorussian + Cantabria + Irish + Kent @ 0.000000
10 Belorussian + Cantabria + Kent + Orcadian @ 0.000000
11 Belorussian + English + Orcadian + Valencia @ 0.000000
12 Aragon + Argyll + Belorussian + British_Isles @ 0.250225
13 Belorussian + English + Orkney + Valencia @ 0.251148
14 Belorussian + British_Isles + Orkney + Valencia @ 0.260586
15 Belorussian + British + Orkney + Valencia @ 0.264416
16 Belorussian + Cantabria + English + Orcadian @ 0.265015
17 Belorussian + British + Orcadian + Spaniards @ 0.269970
18 Belorussian + CEU30 + Orcadian + Valencia @ 0.277117
19 Aragon + Belorussian + CEU30 + Orcadian @ 0.289516
20 Belorussian + Cantabria + Cornwall + Orkney @ 0.300360


HarappaWorld 4-Ancestors Oracle:

1 basque + belorussian + orcadian + slovenian @ 1.011994
2 belorussian + orcadian + slovenian + spain-basc @ 1.071640
3 basque + french + lithuanian + ukranian @ 1.102894
4 hungarian + lithuanian + spain-basc + utahn-white @ 1.127077
5 basque + hungarian + lithuanian + utahn-white @ 1.129016
6 basque + hungarian + lithuanian + n-european @ 1.191291
7 basque + belorussian + slovenian + utahn-white @ 1.208107
8 basque + lithuanian + slovenian + utahn-white @ 1.211915
9 basque + british + ukranian + ukranian @ 1.219691
10 basque + lithuanian + orcadian + serbian @ 1.231788
11 basque + hungarian + lithuanian + utahn-white @ 1.242823
12 basque + orcadian + ukranian + ukranian @ 1.247826
13 belorussian + british + slovenian + spain-basc @ 1.301003
14 lithuanian + slovenian + spain-basc + utahn-white @ 1.382463
15 belorussian + hungarian + orcadian + spain-basc @ 1.389067
16 basque + belorussian + british + slovenian @ 1.405472
17 hungarian + lithuanian + spain-basc + utahn-white @ 1.410242
18 belorussian + slovenian + spain-basc + utahn-white @ 1.414452
19 lithuanian + orcadian + serbian + spain-basc @ 1.441796
20 french + lithuanian + spain-basc + ukranian @ 1.443352


EUtest 4-Ancestors Oracle:

1 Cornish + English + English + English @ 1.230978
2 Cornish + Cornish + English + English @ 1.466153
3 English + English + English + Orcadian @ 1.561817
4 English + English + English + Scottish @ 1.626319
5 Cornish + English + English + Orcadian @ 1.645994
6 English + English + English + IE @ 1.655308
7 English + English + English + English @ 1.703235
8 Cornish + English + English + NL @ 1.747462
9 Cornish + DK + English + English @ 1.803706
10 Cornish + Cornish + DK + English @ 1.817353
11 Cornish + English + English + IE @ 1.882855
12 English + English + NL + Scottish @ 1.896381
13 Cornish + English + English + Scottish @ 1.957002
14 Cornish + Cornish + English + NL @ 1.969627
15 English + English + NL + Orcadian @ 1.981662
16 English + English + IE + NL @ 2.023281
17 English + English + English + NL @ 2.068767
18 Cornish + English + NL + Orcadian @ 2.086267
19 English + FR + NO + Scottish @ 2.128249
20 Cornish + English + English + West_&_Central_German @ 2.168996


Eurogenes EUtest V2 K15 4-Ancestors Oracle:

1 Orcadian + Orcadian + Orcadian + Southwest_English @ 2.801044
2 Orcadian + Orcadian + Southwest_English + Southwest_English @ 2.808933
3 Orcadian + Orcadian + Southwest_English + West_Scottish @ 2.847229
4 Orcadian + Orcadian + Southeast_English + Southwest_English @ 2.934061
5 Irish + Orcadian + Orcadian + Southwest_English @ 2.989665
6 Orcadian + Southwest_English + Southwest_English + West_Scottish @ 3.063306
7 Orcadian + Orcadian + Orcadian + Southeast_English @ 3.134112
8 Orcadian + Southwest_English + West_Scottish + West_Scottish @ 3.159362
9 Orcadian + Orcadian + Southeast_English + Southeast_English @ 3.215602
10 Orcadian + Orcadian + Southeast_English + West_Scottish @ 3.247473
11 Orcadian + Southeast_English + Southwest_English + West_Scottish @ 3.263688
12 Irish + Orcadian + Southwest_English + Southwest_English @ 3.314818
13 Irish + Orcadian + Southwest_English + West_Scottish @ 3.333836
14 Irish + Orcadian + Orcadian + Southeast_English @ 3.344661
15 Orcadian + Orcadian + Orcadian + West_Scottish @ 3.375508
16 Irish + Orcadian + Orcadian + Orcadian @ 3.383590
17 Orcadian + Orcadian + West_Scottish + West_Scottish @ 3.467124
18 Orcadian + Southeast_English + Southwest_English + Southwest_English @ 3.478116
19 North_Dutch + Orcadian + Orcadian + Southwest_English @ 3.483853
20 Irish + Orcadian + Southeast_English + Southwest_English @ 3.484694


Eurogenes K13 4-Ancestors Oracle:

1 Southeast_English + Southwest_English + Southwest_English + West_Scottish @ 2.238837
2 Southeast_English + Southwest_English + West_Scottish + West_Scottish @ 2.261215
3 Southeast_English + Southeast_English + Southwest_English + West_Scottish @ 2.381864
4 Southwest_English + Southwest_English + West_Scottish + West_Scottish @ 2.389854
5 Orcadian + Southwest_English + Southwest_English + West_Scottish @ 2.399714
6 Orcadian + Southeast_English + Southwest_English + West_Scottish @ 2.405423
7 Southwest_English + Southwest_English + Southwest_English + West_Scottish @ 2.413008
8 Southeast_English + Southeast_English + West_Scottish + West_Scottish @ 2.420424
9 Orcadian + Southwest_English + Southwest_English + Southwest_English @ 2.476946
10 Orcadian + Southeast_English + Southwest_English + Southwest_English @ 2.478590
11 Irish + Southeast_English + Southwest_English + Southwest_English @ 2.511645
12 Danish + Southwest_English + Southwest_English + Southwest_English @ 2.516330
13 Irish + Southeast_English + Southwest_English + West_Scottish @ 2.522455
14 Orcadian + Orcadian + Southwest_English + Southwest_English @ 2.548872
15 Southeast_English + Southwest_English + Southwest_English + Southwest_English @ 2.584332
16 Danish + Southwest_English + Southwest_English + West_Scottish @ 2.597174
17 Southeast_English + West_Scottish + West_Scottish + West_Scottish @ 2.597390
18 Irish + Southwest_English + Southwest_English + West_Scottish @ 2.626618
19 North_Dutch + Southwest_English + Southwest_English + Southwest_English @ 2.629165
20 Irish + Southwest_English + Southwest_English + Southwest_English @ 2.633650

Personally, I think Eurogenes' K13, V2 K15 and EUtest (in that order) make the most sense. MDLP K23b is quite good as far as I'm concerned (it's quite accurate, especially since it proxies my father's ancestry as "Cretan" and my mother's ancestry as Norwegian-like), it just gives weird fits for my mother (much like Dodecad V3 and K12b, the latter produces very coherent fits for me though).

MitchellSince1893
01-27-2015, 04:13 PM
I posted this in another thread but here it is again. Center points were derived by plotting out multiple Oracle 4 locations with the smallest genetic distances and getting a center point via http://www.geomidpoint.com/

3611

As you can see, in general Dodecad, MDLP, and SPA give me more southeastern locations than Eurogenes, McDonald and GPS.

You can look at my profile picture and signature decide which is best based on my genealogical and genetic makeup.

I wondered in another thread if Eurogenes K13 and Hunter_Gatherer results are showing the homeland of the Angles, while K15, Dodecade K7b and GPS are showing the destination of the Angles after they crossed the North Sea.

SwampThing27
01-27-2015, 05:17 PM
The Eurogenes tend to give me results that are more likely according to paper, while the Dodecad and MDLP tests tend to give me fst distances much closer to zero. Like fst of 1 compared to fst of 4.

randwulf
01-28-2015, 12:20 AM
I posted this in another thread but here it is again. Center points were derived by plotting out multiple Oracle 4 locations with the smallest genetic distances and getting a center point via http://www.geomidpoint.com/

As you can see, in general Dodecad, MDLP, and SPA give me more southeastern locations than Eurogenes, McDonald and GPS.

You can look at my profile picture and signature decide which is best based on my genealogical and genetic makeup.

I wondered in another thread if Eurogenes K13 and Hunter_Gatherer results are showing the homeland of the Angles, while K15, Dodecade K7b and GPS are showing the destination of the Angles after they crossed the North Sea.

Very interesting. This thread has turned in a direction I was trying to query on another thread. I decided to run with this idea to see if I could find meaning with my four population results. The first thing I did was plot my surnames at the level where I had 32 ancestors, giving each one credit for 1/32 of my ancestry (this is just for some kind of reality check on the plotting). The following map uses the Geo Mid Point link above with the 32 ancestor locations on it. There is one estimation and a few approximations in the 32, but correct as best as I can get them:

3615

Then I did a map like MitchellSince1893's map (not as pretty as his). I put the plot point from my 32 surnames as well as a point that Tolan gave me when he noticed my mix matched his samples from Lillie, France. These two points are only 128 miles/205 km apart, interestingly, so I think they are a pretty good reality check. Then I used the four-population plots like MitchellSince1893 for various calculators. Sorry, that some of the colors are a little hard to read, maybe. It is interesting which ones were closest as well as how far east a couple are pulled, into areas where I have no ancestry of which I know. Fun exercise.

3616

MitchellSince1893
01-28-2015, 03:21 AM
...I did a map like MitchellSince1893's map (not as pretty as his)...

Thanks for the map complement, but all I did was use google earth...turned off a bunch of layers that cluttered things up, and used a basic bullseye symbol to mark each point.

FWIW I went back and created a center point of all the previous admixture tools center points...aka the pointless point. :P Having too much fun with the geographic midpoint calculator

3618

Tolan
01-28-2015, 09:31 PM
I posted this in another thread but here it is again. Center points were derived by plotting out multiple Oracle 4 locations with the smallest genetic distances and getting a center point via http://www.geomidpoint.com/

3611

As you can see, in general Dodecad, MDLP, and SPA give me more southeastern locations than Eurogenes, McDonald and GPS.

You can look at my profile picture and signature decide which is best based on my genealogical and genetic makeup.

I wondered in another thread if Eurogenes K13 and Hunter_Gatherer results are showing the homeland of the Angles, while K15, Dodecade K7b and GPS are showing the destination of the Angles after they crossed the North Sea.

Until now, I was not paying attention to Oracle 4 populations: results seemed strange to me!
But by averaging the four populations, there is something coherent!
Thanks MitchellSince1893 for giving the link to this website to calculate the average! ;)

So I did the same with several calculators + both positions by SPA (in blue):
3619


And the average of 8 calculators:
3620

randwulf
02-01-2015, 07:11 PM
I was rereading this thread (too much time on my hands?) and noticed MitchellSince1893 had also plotted MDLP K5 and MDLP K7 with "midpoints" of the four population results. I had ignored those in my earlier post, but ran them to see what happened. I got four "0.00" distance results from MDLP K5 and plotted all four. I got very accurate results from all, especially three of them which were closer to my two speculative "actual" points than any of the other calculators.

3692

Krissie81
07-11-2016, 04:07 AM
For GEDmatch which admixture would be best for me?
I am:
Europe: 87%
Italy/Greece 56%
Ireland 10%
Europe East 9%
Great Britain 6%
European Jewish 2%
Iberian Peninsula 2%
Finland/Northwest Russia 2%
Middle East 8%
Caucasus 5%

Krissie81
07-11-2016, 04:08 AM
For GEDmatch which admixture would be best for me?
I am:
Europe: 87%
Italy/Greece 56%
Ireland 10%
Europe East 9%
Great Britain 6%
European Jewish 2%
Iberian Peninsula 2%
Finland/Northwest Russia 2%
Middle East 8%
Caucasus 5%

wombatofthenorth
07-11-2016, 05:04 AM
At present, for West Eurasians, I'd consider the MDLP K23b to be the most informative calculator for two reasons:


If you use FTDNA/Geno 2.0 NG transferred data then MDLP K23b does seem pretty good!
My mom got:
Ancestral_Altaic 1.62
Arctic 1.02
Caucasian 26.95
Archaic_Human 0.10
European_Early_Farmers 5.30
Melano_Polynesian 0.93
Archaic_African 0.31
Near_East 0.95
European_Hunters_Gatherers 62.83

Using 1 population approximation:
1 Lithuanian @ 3.064577
2 Latvian @ 3.872536
3 Estonian @ 4.899636
4 Belarusian_South @ 5.990172

Using 2 populations approximation:
1 50% Lithuanian +50% Lithuanian @ 3.064577

Using 3 populations approximation:
1 50% Lithuanian +25% Lithuanian +25% Lithuanian @ 3.064577

On paper her:

mother's father's side is all traced in Latvia to a single parish into the 1700 and 1600s (as far back as 1649);

her father's father's side is all trace to a single Parish in Latvia to 1800 and mostly 1700s (as far back as 1737);

her father's mother's side is mostly all traced back to a single Parish in Latvia to the 1800 and 1700s (as far back as around 1740) although one line does have no father listed so it's a bit unsure but she has some genetic matches to some who seem to be mostly Latvians who lived over the border in Lithuania and one line that apparently heads back to France from there (a few matches seem to get slightly elevated Oceanian on Eurogenes K13 as does my mom);

her mother's mother's line is confusing, can't trace it yet, seemingly German but her ancestry tests, all of them, show way less than 25% typical German, not even 12.5%, maybe it's some sort of very eastern German or Baltic German that had lots of Baltic blood mixed in or some bits from Eastern Europe or something or maybe each parent on that too was only half German and half Latvian too or something, not sure.

anyway it seems like at least 73% of her background is Baltic going back hundreds of years and maybe more like 86%.

Maybe the oracles give Lithuania instead of Latvian because the bit of German pulls it overall a touch south.

-------------------
using 23andme raw data her oracles are fairly different though (the 1 pop and 2 pop seem much better with the Geno 2.0 NG data while the 3 pop depends on how much German there is and if maybe it gets called Polish although this results with 75% Lith and 25% Polish seems a bit too far pulled south overall and the Geno 2.0 NG data doesn't pull her result as far south although OTOH it doesn't directly hint at German/Polish; for whatever reason this Oracle 4 always times out on her data before it manages to complete the 4 pop answer :( ):
Using 1 population approximation:
1 Lithuanian @ 3.095196
2 Belarusian_Russian @ 4.466355
3 Mixed_East_Slav @ 4.511575
4 Latvian @ 4.592929

Using 2 populations approximation:
1 50% Belarusian_Russian +50% Lithuanian @ 2.663472

Using 3 populations approximation:
1 50% Lithuanian +25% Lithuanian +25% Pole @ 2.516112
-------------------------------------------

my dad's results seem a little weird though (on paper all Latvian so far, although we can't track it back nearly as well; his ancestry tests from major companies hint at Latvian plus probably one Finnish ancestor within typical family tree time frame and probably 1 or maybe 2 of the same but from maybe around Romania or the western Ukraine or something, talking maybe 3rd to 4th great-grandparent level in each case:
23andme data:

Using 1 population approximation:
1 Russian-North-West @ 4.635733
2 Sorb @ 5.945943
3 Russian-West @ 6.192382

Using 2 populations approximation:
1 50% German +50% Latvian @ 2.628950

Using 3 populations approximation:
1 50% Balt +25% Latvian +25% Norwegian_West @ 1.285094

--------
and with Geno 2.0 NG data:
Using 1 population approximation:
1 Russian-North-West @ 4.100970
2 Balt @ 5.138099
3 Russian-West @ 5.848875
4 Sorb @ 5.858302

Using 2 populations approximation:
1 50% Estonian +50% Kashub @ 2.316044

Using 3 populations approximation:
1 50% Latvian +25% Pole +25% Swede @ 1.651963

I feel like the Oracles 1 pop and 2 pop are definitely more accurate using Geno 2.0 NG data than 23andme while with the 3 pop it's a little harder to say yet.
------------------------------------------------

Eurogenes K13 gives me (23andme data):
1 Lithuanian @ 2.595495
2 Belorussian @ 5.072916
3 Estonian @ 5.450169
4 Estonian_Polish @ 5.524313
Using 2 populations approximation:
1 50% Lithuanian +50% Lithuanian @ 2.595495

Using 3 populations approximation:
1 50% Lithuanian +25% Lithuanian +25% Polish @ 2.179320

Using 4 populations approximation:
1 Lithuanian + Lithuanian + Lithuanian + Polish @ 2.179320
2 Belorussian + Lithuanian + Lithuanian + Lithuanian @ 2.266166
3 Estonian + Lithuanian + Lithuanian + Lithuanian @ 2.272408

and my mom (23andme data):
Using 1 population approximation:
1 Lithuanian @ 6.115575
2 Estonian @ 7.558150
3 Belorussian @ 8.781824
4 Estonian_Polish @ 10.013271

Using 2 populations approximation:
1 50% Lithuanian +50% Lithuanian @ 6.115575

Using 3 populations approximation:
1 50% Lithuanian +25% Lithuanian +25% Lithuanian @ 6.115575

Using 4 populations approximation:
1 Estonian + Lithuanian + Lithuanian + Lithuanian @ 5.914726 (interesting result in that half her ancestry is from the far southern edge of Latvia on the Lithuanian border and at least half of the other half from the northern part of Latvia getting closer to Estonia)
2 Lithuanian + Lithuanian + Lithuanian + Lithuanian @ 6.115575
3 Estonian + Estonian + Lithuanian + Lithuanian @ 6.124651

my mom (using Geno 2.0 NG data):
Using 1 population approximation:
1 Lithuanian @ 5.697521
2 Estonian @ 5.996624
3 Belorussian @ 7.218196
4 Estonian_Polish @ 8.583775

Using 2 populations approximation:
1 50% Estonian +50% Lithuanian @ 4.948126

Using 3 populations approximation:
1 50% Estonian +25% Lithuanian +25% Lithuanian @ 4.948126

Using 4 populations approximation:
1 Estonian + Estonian + Lithuanian + Lithuanian @ 4.948126
2 Lithuanian + Lithuanian + Lithuanian + Southwest_Finnish @ 5.096645
3 Estonian + Lithuanian + Lithuanian + Lithuanian @ 5.103139
4 Belorussian + Estonian + Lithuanian + Lithuanian @ 5.133492
--------------------------
my dad on Eurogenes K13 (23andme data):
Using 1 population approximation:
1 Lithuanian @ 4.344326
2 Russian_Smolensk @ 5.067872
3 Estonian_Polish @ 5.763341
4 Belorussian @ 6.288128

Using 2 populations approximation:
1 50% Lithuanian +50% Russian_Smolensk @ 3.407391

Using 3 populations approximation:
1 50% Lithuanian +25% Lithuanian +25% Ukrainian_Lviv @ 3.251049

Using 4 populations approximation:
1 Lithuanian + Lithuanian + Lithuanian + Ukrainian_Lviv @ 3.251049
2 Lithuanian + Lithuanian + Lithuanian + Ukrainian @ 3.304688
3 Lithuanian + Lithuanian + Lithuanian + Ukrainian_Belgorod @ 3.347152
4 Lithuanian + Lithuanian + Lithuanian + Polish @ 3.393333
-----
and my dad (Geno 2.0 NG data):
Using 1 population approximation:
1 Lithuanian @ 4.722293
2 Russian_Smolensk @ 4.795887
3 Estonian_Polish @ 5.649219
4 Belorussian @ 6.218225

Using 2 populations approximation:
1 50% Lithuanian +50% Russian_Smolensk @ 3.461540

Using 3 populations approximation:
1 50% Lithuanian +25% Lithuanian +25% Ukrainian_Lviv @ 3.258453

Using 4 populations approximation:
1 Lithuanian + Lithuanian + Lithuanian + Ukrainian_Lviv @ 3.258453
2 Lithuanian + Lithuanian + Polish + Ukrainian_Belgorod @ 3.392436
3 Lithuanian + Lithuanian + Lithuanian + Ukrainian @ 3.422159
4 Lithuanian + Lithuanian + Russian_Smolensk + Ukrainian_Belgorod @ 3.455401
--------------------------------------------
Looking at Oracle 4 One Pop mode the one that gives all of us as the #1 choice the country right next to our country where most of our ancestors come from is Eurogenes K13. MDLP K23b does it for my mom but not my dad.

Missouri1455
11-10-2016, 04:28 PM
What is the best program to use for a middle eastern with a minor east african admixture?

Underworld
11-10-2016, 07:11 PM
With gedmatch calculators, accuracy comes in different ways...

From what I've observed, MDLP k23b seems to be the best for the Oracle, together with EUTest. MDLP k23b's Oracle looks best when calculating for Southern Europeans, not so much for Northern Europeans, for which EUTest is probably the best.

For the actual percentage of the components, EUTest (the first one, not the k15v2) seems again to be the one with the least calculation errors and best percentages (mostly for European though...).

Eurogenes K36 is IMO the best for studying deep ancestries of populations - what each population is made of on a deeper level, although it's hard to interpret the results without the required knowledge.

Underworld
11-10-2016, 07:15 PM
What is the best program to use for a middle eastern with a minor east african admixture?

I'd try Dodecad K12b and Eurogenes K36 :)

Dodecad K12b distinguishes between Southwest Asian, Northwest African, Caucasus, East African, West African, South Asian, etc. Seems to have all the necessary components to give you a good reading of your ancestry. Same thing with Eurogenes K36.

kikkk
11-10-2016, 08:08 PM
What is the best program to use for a middle eastern with a minor east african admixture?

In my opinion, most of calcs are good, but the "easiest" ones to interpret are the ones without ancient samples and with a moderate number of components (preferably between 8 and 16; as below 8 may not be very informative for some sub-components and above 16 could be mind-baffling for beginners).
Also there are some very specific ones (like all Africa and Ethiohelix Japanese) that are not very informative if your aim is merely to get a clear picture of your overall autosomal makeup.
Hence, calculators such as Dodecad v3, Eurogenes k9=>k15 and puntdnal k12 modern&k15 as well as MDLP world are very informaive IMO.

JerryS.
11-11-2017, 06:53 PM
late to the party, but I've found that using the Oracle 4 populations for any calculating model to be misleading. I say this because they only go as low as 25% for any given population, meaning if you are 15% Greek you might get pigeon holed as 25% Romanian instead... unless you have at least 1/4 of whatever group you know of, the regular Oracle using only 2 population groupings seems more likely to be accurate regarding your minor ethnicity group.

Kqokanm
11-19-2017, 01:53 AM
MDLP k23b imo

JerryS.
11-19-2017, 04:53 AM
MDLP k23b imo

I agree regarding the regular oracle primary and secondary populations (that and Dodecad V3). however, as with any oracle 4, if you're not at least 25% of something you will get pigeon holed into something close.

kevinduffy
11-19-2017, 05:06 AM
For me the best one is puntDNAL K13 Global.

Teutorigos
09-01-2018, 09:41 AM
I am new to this forum and the concept of DNA Ancestry, i have only turned to it recently when my paper trails hit dead ends.

Please can anything suggest which out of Eurogenes, MDLP and Dodecad is the most accurate or reliable based on results that make sense to members own Ethnic Ancestry?

Also whether the Oracle or Oracle-4 is better than the other, i seem to gain crazy results that do not add up on the Gedmatch basic oracle but better hits on the Aleksandr Burnashev version.

All opinions professional or otherwise would be of great help.

puntDNAL K12 Modern Oracle, in my humble opinion, but I ignore everything below the single population number 1 oracle result because we British are just super-sophisticated like that :

https://anthrogenica.com/showthread.php?6400-puntDNAL-K12-Modern&p=476781#post476781

PoxVoldius
09-01-2018, 05:32 PM
For single populations, with my family's mixed heritage (mostly mixed German populations on Dad's side, lots of Colonial American + 1/4 Czech on Mom's side), I would say that Dodecad World9, PuntDNAL K15, and Dodecad K7b probably got the closest.

Hard to say on mixed-mode, 2-pop, 3-pop & 4-pop, though. We're mixed enough to confuse the oracles there, I think.

KingKhalasi
09-02-2018, 09:03 AM
I don't think there is a "Most accurate" or "best calculator" on Gedmatch, each test has it's own purpose and each test has it's areas it's better at than another calc. For example, EthioHelix does a way better job for africans and African admixed people like me than any other calculator available. Dodecad Africa9 isn't too bad either, but i'd still prefer the EthioHelix K10 + French because it's easier to read it's results.


My personal best tests are from best to worst:

EthioHelix > Gedrosia > Dodecad > MDLP > Others


and my best test outside from Gedmatch is MyGenomeBox, mainly because it gets very close to my expected percentages I should inherit.


So as you can see, tests are different for everybody, EthioHelix is accurate for me, but it will never be accurate for a full brit, or euromutt for example.

JerryS.
09-02-2018, 08:27 PM
I don't think there is a "Most accurate" or "best calculator" on Gedmatch, each test has it's own purpose and each test has it's areas it's better at than another calc. For example, EthioHelix does a way better job for africans and African admixed people like me than any other calculator available. Dodecad Africa9 isn't too bad either, but i'd still prefer the EthioHelix K10 + French because it's easier to read it's results.


My personal best tests are from best to worst:

EthioHelix > Gedrosia > Dodecad > MDLP > Others


and my best test outside from Gedmatch is MyGenomeBox, mainly because it gets very close to my expected percentages I should inherit.


So as you can see, tests are different for everybody, EthioHelix is accurate for me, but it will never be accurate for a full brit, or euromutt for example.

absolutely.

for people that are homogeneous there are quite a few that will work for them and most others like them. but for folks of mixed regions like me, or even mixed races like some others..... the percentage of this or that which you have will determine which model best reflects your mix.

KingKhalasi
09-02-2018, 08:34 PM
absolutely.

for people that are homogeneous there are quite a few that will work for them and most others like them. but for folks of mixed regions like me, or even mixed races like some others..... the percentage of this or that which you have will determine which model best reflects your mix.

Jup indeed XD my best tests are Dodecad (excluding Dodecad V3), EthioHelix, and Gedrosia. A test that's average for me, not good or bad, are MDLP Calculators. Tests that are least good for me are HarappaWorld, Eurogenes, and puntDNAL. especially eurogenes EUtest, Eurogenes K15, Eurogenes K13, and K36. and basically every other test of eurogenes but these ones are by far the worst, especially K36 takes the cake.


EthioHelix perfectly balances my ethnicities, check out my oracles:


# Population Percent
1 French 58.69
2 West-Africa 20.55
3 North-Africa 14.08
4 Eastern-Bantu 2.85
5 Biaka-Pygmy 1.99
6 Omotic 0.87
7 Nilo-Saharan 0.85
8 Khoi-San 0.11

Least-squares method.

Using 1 population approximation:
1 egypt @ 28.546587
2 egyptans @ 29.966671
3 libya @ 37.318996
4 french @ 48.370476
5 algeria @ 51.961971
6 morocco-n @ 53.463043
7 EtA @ 59.000324
8 EtT @ 59.366028
9 ethiopian-jews @ 59.730007
10 moroccans @ 60.164871
11 morocco-s @ 62.756523
12 EtT-P @ 63.916306
13 EtA-P @ 66.361290
14 EtO-P @ 67.930603
15 Afar @ 67.959602
16 fulani @ 68.008682
17 EtO @ 69.372124
18 hema @ 69.376450
19 fang @ 73.513809
20 sahara-occ @ 73.785797

Using 2 populations approximation:
1 50% french +50% fulani @ 10.893702


Using 3 populations approximation:
1 50% french +25% morocco-n +25% yoruba @ 2.964315


Using 4 populations approximation:
+++++++++++++++++++++++++++++++++++
1 algeria + bambaran + french + french @ 1.882089
2 bambaran + french + french + morocco-n @ 1.951400
3 algeria + brong + french + french @ 2.032186
4 brong + french + french + morocco-n @ 2.141153
5 algeria + french + french + yoruba @ 2.832829
6 algeria + french + french + yoruba @ 2.832829
7 brong + french + french + moroccans @ 2.857018
8 french + french + morocco-n + yoruba @ 2.964315
9 french + french + morocco-n + yoruba @ 2.964315
10 bambaran + french + french + moroccans @ 3.036810
11 french + french + moroccans + yoruba @ 3.471935
12 french + french + moroccans + yoruba @ 3.471935
13 dogon + french + french + morocco-n @ 3.547430
14 algeria + french + french + mandenka @ 3.594868
15 algeria + dogon + french + french @ 3.670599
16 french + french + mandenka + morocco-n @ 3.691630
17 dogon + french + french + moroccans @ 4.452665
18 algeria + french + french + hausa @ 4.537116
19 french + french + mandenka + moroccans @ 4.539755
20 french + french + hausa + morocco-n @ 4.726855

Mixed Mode Population Sharing:

# Primary Population (source) Secondary Population (source) Distance
1 58.5% french + 41.5% fulani @ 4.33
2 60.6% french + 39.4% kaba @ 15.99
3 57.2% french + 42.8% morocco-s @ 16.03
4 63% french + 37% mandenka @ 16.52
5 61.8% french + 38.2% hausa @ 16.57
6 61.8% french + 38.2% igbo @ 16.65
7 63% french + 37% bambaran @ 16.81
8 61% french + 39% bamoun @ 16.97
9 62.7% french + 37.3% yoruba @ 17.05
10 62.8% french + 37.2% brong @ 17.17
11 60.6% french + 39.4% fang @ 17.97
12 60.9% french + 39.1% kongo @ 18.09
13 60.8% french + 39.2% mada @ 18.19
14 64.3% french + 35.7% dogon @ 18.29
15 82.4% egypt + 17.6% brong @ 19.54
16 83.4% egypt + 16.6% dogon @ 19.61
17 82.7% egypt + 17.3% bambaran @ 19.69
18 82.6% egypt + 17.4% yoruba @ 19.74
19 82.9% egypt + 17.1% mandenka @ 19.9
20 82.4% egypt + 17.6% hausa @ 20.12

JerryS.
09-02-2018, 08:38 PM
Jup indeed XD my best tests are Dodecad (excluding Dodecad V3), EthioHelix, and Gedrosia. A test that's average for me, not good or bad, are MDLP Calculators. Tests that are least good for me are HarappaWorld, Eurogenes, and puntDNAL. especially eurogenes EUtest, Eurogenes K15, Eurogenes K13, and K36. and basically every other test of eurogenes but these ones are by far the worst, especially K36 takes the cake.


EthioHelix perfectly balances my ethnicities, check out my oracles:


# Population Percent
1 French 58.69
2 West-Africa 20.55
3 North-Africa 14.08
4 Eastern-Bantu 2.85
5 Biaka-Pygmy 1.99
6 Omotic 0.87
7 Nilo-Saharan 0.85
8 Khoi-San 0.11

Least-squares method.

Using 1 population approximation:
1 egypt @ 28.546587
2 egyptans @ 29.966671
3 libya @ 37.318996
4 french @ 48.370476
5 algeria @ 51.961971
6 morocco-n @ 53.463043
7 EtA @ 59.000324
8 EtT @ 59.366028
9 ethiopian-jews @ 59.730007
10 moroccans @ 60.164871
11 morocco-s @ 62.756523
12 EtT-P @ 63.916306
13 EtA-P @ 66.361290
14 EtO-P @ 67.930603
15 Afar @ 67.959602
16 fulani @ 68.008682
17 EtO @ 69.372124
18 hema @ 69.376450
19 fang @ 73.513809
20 sahara-occ @ 73.785797

Using 2 populations approximation:
1 50% french +50% fulani @ 10.893702


Using 3 populations approximation:
1 50% french +25% morocco-n +25% yoruba @ 2.964315


Using 4 populations approximation:
+++++++++++++++++++++++++++++++++++
1 algeria + bambaran + french + french @ 1.882089
2 bambaran + french + french + morocco-n @ 1.951400
3 algeria + brong + french + french @ 2.032186
4 brong + french + french + morocco-n @ 2.141153
5 algeria + french + french + yoruba @ 2.832829
6 algeria + french + french + yoruba @ 2.832829
7 brong + french + french + moroccans @ 2.857018
8 french + french + morocco-n + yoruba @ 2.964315
9 french + french + morocco-n + yoruba @ 2.964315
10 bambaran + french + french + moroccans @ 3.036810
11 french + french + moroccans + yoruba @ 3.471935
12 french + french + moroccans + yoruba @ 3.471935
13 dogon + french + french + morocco-n @ 3.547430
14 algeria + french + french + mandenka @ 3.594868
15 algeria + dogon + french + french @ 3.670599
16 french + french + mandenka + morocco-n @ 3.691630
17 dogon + french + french + moroccans @ 4.452665
18 algeria + french + french + hausa @ 4.537116
19 french + french + mandenka + moroccans @ 4.539755
20 french + french + hausa + morocco-n @ 4.726855

Mixed Mode Population Sharing:

# Primary Population (source) Secondary Population (source) Distance
1 58.5% french + 41.5% fulani @ 4.33
2 60.6% french + 39.4% kaba @ 15.99
3 57.2% french + 42.8% morocco-s @ 16.03
4 63% french + 37% mandenka @ 16.52
5 61.8% french + 38.2% hausa @ 16.57
6 61.8% french + 38.2% igbo @ 16.65
7 63% french + 37% bambaran @ 16.81
8 61% french + 39% bamoun @ 16.97
9 62.7% french + 37.3% yoruba @ 17.05
10 62.8% french + 37.2% brong @ 17.17
11 60.6% french + 39.4% fang @ 17.97
12 60.9% french + 39.1% kongo @ 18.09
13 60.8% french + 39.2% mada @ 18.19
14 64.3% french + 35.7% dogon @ 18.29
15 82.4% egypt + 17.6% brong @ 19.54
16 83.4% egypt + 16.6% dogon @ 19.61
17 82.7% egypt + 17.3% bambaran @ 19.69
18 82.6% egypt + 17.4% yoruba @ 19.74
19 82.9% egypt + 17.1% mandenka @ 19.9
20 82.4% egypt + 17.6% hausa @ 20.12

yes, with a considerable sub-Sahara African background you really have to choose the right model. Eurogenes is probably the worst for me when it comes to showing my Mediterranean background, but Dodecad and MDLP seem to hit it reasonably well. of course Eurogenes works best for deciphering my N.W. European ethnicities.

KingKhalasi
09-02-2018, 08:43 PM
yes, with a considerable sub-Sahara African background you really have to choose the right model. Eurogenes is probably the worst for me when it comes to showing my Mediterranean background, but Dodecad and MDLP seem to hit it reasonably well. of course Eurogenes works best for deciphering my N.W. European ethnicities.

Lol, funnily enough barely any test is good for my European side, I can only go by paper trail and genealogy because calculators aint telling me much. One time im scoring irish/Hungarian, the other time german and british, then the other time Scottish and Norwegian, then another time "canary islander" "Tuscan" or "Andalusia" meanwhile I don't even have Mediterranean ancestry XDD


eitherway yeah, then again EthioHelix is for "African" admixed people, not necessarily SSA I think. They also have a cluster for North Africa, which of everybody who takes the test scores some of, but obviously for Europeans it means nothing and is just excess European or Mediterranean.

In my case I am half euro yet I score equal amounts of North African as full euros XD eitherway, oracles are low, I get Algeria/SSA + 50% euro for my 4th pop oracles so its accurate I think. However I think im more Oromo Ethiopian rather than Berber but whatever XD close enough

JerryS.
09-02-2018, 08:53 PM
Lol, funnily enough barely any test is good for my European side, I can only go by paper trail and genealogy because calculators aint telling me much. One time im scoring irish/Hungarian, the other time german and british, then the other time Scottish and Norwegian, then another time "canary islander" "Tuscan" or "Andalusia" meanwhile I don't even have Mediterranean ancestry XDD


eitherway yeah, then again EthioHelix is for "African" admixed people, not necessarily SSA I think. They also have a cluster for North Africa, which of everybody who takes the test scores some of, but obviously for Europeans it means nothing and is just excess European or Mediterranean.

In my case I am half euro yet I score equal amounts of North African as full euros XD eitherway, oracles are low, I get Algeria/SSA + 50% euro for my 4th pop oracles so its accurate I think. However I think im more Oromo Ethiopian rather than Berber but whatever XD close enough

yes, as the name would indicate (Euro)genes, anything that has a mix from other regions tends to give it difficulty. Mediterranean groups have MENA more or less depending on which group and that tends to give it fits when the amount is less than 25% because it then tries to pigeon hole it somewhere else... in your case, having more than 25% from another continent gives Eurogenes difficulty because it isn't trying to place each group into its proper place, its trying to combined them to find a good fit.

KingKhalasi
09-02-2018, 09:45 PM
yes, as the name would indicate (Euro)genes, anything that has a mix from other regions tends to give it difficulty. Mediterranean groups have MENA more or less depending on which group and that tends to give it fits when the amount is less than 25% because it then tries to pigeon hole it somewhere else... in your case, having more than 25% from another continent gives Eurogenes difficulty because it isn't trying to place each group into its proper place, its trying to combined them to find a good fit.

That's true XD im not super good in intepretting results but what I do know is that theres no way in hell that I am 65%> euro. that's all im sayin XD if MyGenomeBox is good in general percentages (not specific) and it gives me 54.14% euro, along many other tests im pretty sure its true.

only Wegene, AncestryDNA, and Myheritage give me high European. DNA.LAND maybe, idk if it counts. it gives me 3.2% finnish, 2.2% Balkan, and 13% Mediterranean islander, of which are all part of my Ethiopian/berber ancestry cuz it only gave me 21% African too. so it basically took 17% from my African and just added it to Europe. LOL.

eitherway yeah, eurogenes is best for Europeans and homogenous ppls ofcourse. I expected that before, but ppl said I was 71% euro and 29% SSA just at the hand of my Mixed Oracle mode results in Eurogenes K13, it gave me 71% Danish and 29% SSA I think. well no shit, it adds literally ALL percentages like west Asian, red sea, Mediterranean to the European percentage XD meanwhile it shouldn't even be added to my euro. Doug McDonald and many other people said I had atleast 10-14% ancestry from the Maghreb/Egypt region, including Mephisto from this forum.

Kaspias
09-02-2018, 10:19 PM
That's true XD im not super good in intepretting results but what I do know is that theres no way in hell that I am 65%> euro. that's all im sayin XD if MyGenomeBox is good in general percentages (not specific) and it gives me 54.14% euro, along many other tests im pretty sure its true.

only Wegene, AncestryDNA, and Myheritage give me high European. DNA.LAND maybe, idk if it counts. it gives me 3.2% finnish, 2.2% Balkan, and 13% Mediterranean islander, of which are all part of my Ethiopian/berber ancestry cuz it only gave me 21% African too. so it basically took 17% from my African and just added it to Europe. LOL.

eitherway yeah, eurogenes is best for Europeans and homogenous ppls ofcourse. I expected that before, but ppl said I was 71% euro and 29% SSA just at the hand of my Mixed Oracle mode results in Eurogenes K13, it gave me 71% Danish and 29% SSA I think. well no shit, it adds literally ALL percentages like west Asian, red sea, Mediterranean to the European percentage XD meanwhile it shouldn't even be added to my euro. Doug McDonald and many other people said I had atleast 10-14% ancestry from the Maghreb/Egypt region, including Mephisto from this forum.

hi yemeni

KingKhalasi
09-02-2018, 10:31 PM
hi yemeni

helloo may bruddah XD

theapricity is full of dicks lmao so im glad they banned me, this is a way better side tbh. Maybe less active/popular but still more useful responses when you do get them XD

PoxVoldius
09-03-2018, 02:27 AM
Lol, funnily enough barely any test is good for my European side, I can only go by paper trail and genealogy because calculators aint telling me much. One time im scoring irish/Hungarian, the other time german and british, then the other time Scottish and Norwegian, then another time "canary islander" "Tuscan" or "Andalusia" meanwhile I don't even have Mediterranean ancestry XDD


eitherway yeah, then again EthioHelix is for "African" admixed people, not necessarily SSA I think. They also have a cluster for North Africa, which of everybody who takes the test scores some of, but obviously for Europeans it means nothing and is just excess European or Mediterranean.

In my case I am half euro yet I score equal amounts of North African as full euros XD eitherway, oracles are low, I get Algeria/SSA + 50% euro for my 4th pop oracles so its accurate I think. However I think im more Oromo Ethiopian rather than Berber but whatever XD close enough

Some of the calculator oracles do okay for me on single populations, but they give me all kinds of wacky things in mixed mode.

My shortest distance mixed mode result is from Dodecad K7b, which gives me 77.7% Polish + 22.3% Canarias, at a distance of 1.07. And the most recent possible source of Mediterranean or Iberian in my paper trail is a French Canadian 4th-great-grandfather.

There are 8 different calculators that have given me mixed mode combinations of either NW or NE Europe + Spanish or Portuguese.

I had one result from MDLP World that was 75.7% Slovakian + 24.3% Romania -- not even close to my paper trail.
I had another one from Dodecad K7b that tried to give me 90.6% Norwegian + 9.4% Yemenese.
I had one from Eurogenes K13 that came up as 90.8% Swedish + 9.2% Algerian Jewish.
Eurogenes K15 gave me one that was 73.1% Irish + 26.9% Ukrainian.
PuntDNAL K13 gave me one that was 97.6% Utahn European + 2.4% Cambodian.

Although, one of my mixed mode results from puntDNAL K12 Modern is 98.4% Utahn European + 1.6% African American, at a distance of 2.02, which might actually be a fairly reasonable model for me. (My 5th chromosome from my dad's side has a good sized chunk that consistently comes up as some kind of African in various chromosome browsers, including the one at 23andMe.)

KingKhalasi
09-03-2018, 04:30 AM
Some of the calculator oracles do okay for me on single populations, but they give me all kinds of wacky things in mixed mode.

My shortest distance mixed mode result is from Dodecad K7b, which gives me 77.7% Polish + 22.3% Canarias, at a distance of 1.07. And the most recent possible source of Mediterranean or Iberian in my paper trail is a French Canadian 4th-great-grandfather.

There are 8 different calculators that have given me mixed mode combinations of either NW or NE Europe + Spanish or Portuguese.

I had one result from MDLP World that was 75.7% Slovakian + 24.3% Romania -- not even close to my paper trail.
I had another one from Dodecad K7b that tried to give me 90.6% Norwegian + 9.4% Yemenese.
I had one from Eurogenes K13 that came up as 90.8% Swedish + 9.2% Algerian Jewish.
Eurogenes K15 gave me one that was 73.1% Irish + 26.9% Ukrainian.
PuntDNAL K13 gave me one that was 97.6% Utahn European + 2.4% Cambodian.

Although, one of my mixed mode results from puntDNAL K12 Modern is 98.4% Utahn European + 1.6% African American, at a distance of 2.02, which might actually be a fairly reasonable model for me. (My 5th chromosome from my dad's side has a good sized chunk that consistently comes up as some kind of African in various chromosome browsers, including the one at 23andMe.)


Lol nice XD maybe the African came from slavery of sorts, it might not be much but its definitely there XD


And lmao, my only good mixed mode oracles are those of my EthioHelix K10 + French. you can clearly see its more realistic for my breakdown because the mixed oracle distances are like @1 point lower than those in eurogenes, and the 4th pop approx. are like 3x as accurate XD I have oracles of @6-7 oracle distance on eurogenes 4th pop approx. XDD quite sad tbh. meanwhile ethiohelix gives me oracle distances of @1.8 all the way to @3 so its way more accurate XD

PoxVoldius
09-03-2018, 05:55 AM
Lol nice XD maybe the African came from slavery of sorts, it might not be much but its definitely there XD


And lmao, my only good mixed mode oracles are those of my EthioHelix K10 + French. you can clearly see its more realistic for my breakdown because the mixed oracle distances are like @1 point lower than those in eurogenes, and the 4th pop approx. are like 3x as accurate XD I have oracles of @6-7 oracle distance on eurogenes 4th pop approx. XDD quite sad tbh. meanwhile ethiohelix gives me oracle distances of @1.8 all the way to @3 so its way more accurate XD

Based on DNA matches, I *think* I have that bit of DNA narrowed down to the French Canadian 4th-great-grandfather. Interestingly, all the records I can find for him come from after he and his 2nd wife & kids took up residence in Ontario. Haven't been able to find out who his parents were. And his name doesn't exactly strike me as French -- David Wright -- but he consistently claimed French Canadian from Quebec on the census.

My lowest distance on 4-pop Eurogenes is 1.73 on the EUtest -- English + French + Lithuanian + West & Central German.
My dad's lowest distance on 4-pop Eurogenes is 2.94 on K13 -- Belorussian + Hungarian + North Dutch + Southeast English.

A lot of my 4-pop oracle results want to throw in things like Polish, Belorussian, Ukrainian, Russian, Estonian, Lithuanian, Norwegian, Swedish, and Spanish.

I suppose the closest one to the paper trail is 3.19 on puntDNAL K13 Global -- German North + Utahn European x3.

Single-pop with Eurogenes, my lowest distance is on the K15, North German at 4.95.
Dad's lowest single-pop distance with Eurogenes is on K13, East German at 6.13.

KingKhalasi
09-03-2018, 06:06 AM
Based on DNA matches, I *think* I have that bit of DNA narrowed down to the French Canadian 4th-great-grandfather. Interestingly, all the records I can find for him come from after he and his 2nd wife & kids took up residence in Ontario. Haven't been able to find out who his parents were. And his name doesn't exactly strike me as French -- David Wright -- but he consistently claimed French Canadian from Quebec on the census.

My lowest distance on 4-pop Eurogenes is 1.73 on the EUtest -- English + French + Lithuanian + West & Central German.
My dad's lowest distance on 4-pop Eurogenes is 2.94 on K13 -- Belorussian + Hungarian + North Dutch + Southeast English.

A lot of my 4-pop oracle results want to throw in things like Polish, Belorussian, Ukrainian, Russian, Estonian, Lithuanian, Norwegian, Swedish, and Spanish.

I suppose the closest one to the paper trail is 3.19 on puntDNAL K13 Global -- German North + Utahn European x3.

Single-pop with Eurogenes, my lowest distance is on the K15, North German at 4.95.
Dad's lowest single-pop distance with Eurogenes is on K13, East German at 6.13.

That's cool haha XD on Gedrosia K3 my lowest 1st pop oracle ever was for Yemen surprisingly XD a oracle distance of @6.96 which is fucking close for a person like me, all my 1st pops are always in the @28 distance area, my second lowest is Moroccan at a distance of @19 or @16 but yeah, im mixed as hell XD

PoxVoldius
09-03-2018, 06:18 AM
That's cool haha XD on Gedrosia K3 my lowest 1st pop oracle ever was for Yemen surprisingly XD a oracle distance of @6.96 which is fucking close for a person like me, all my 1st pops are always in the @28 distance area, my second lowest is Moroccan at a distance of @19 or @16 but yeah, im mixed as hell XD

Gedrosia K3 gives some pretty strange oracle results, probably from only having 3 very broad components it's looking at. For single-pop it gave me a 0.00-distance result of Georgian, and for 3-pop it gave me a 0.00-distance result of Abkhasian + Albanian + Cypriot.

KingKhalasi
09-03-2018, 06:43 AM
Gedrosia K3 gives some pretty strange oracle results, probably from only having 3 very broad components it's looking at. For single-pop it gave me a 0.00-distance result of Georgian, and for 3-pop it gave me a 0.00-distance result of Abkhasian + Albanian + Cypriot.

Well yeah gedrosia K3 is weird as fuck XD I think its only accurate for MENA's and MENA admixed people

alexfritz
09-03-2018, 03:57 PM
i actually like all the calcs on GEDmatch each informative on its own, but for favorite my opinion is they are not on GEDmatch but on yourDNAportal maybe since i used everything one can do with V2K15 and K36 i like alot the K47 lukaszM and also the K33 MDLP would be cool if you could do as much with the K47 as with the K36;

KingKhalasi
09-03-2018, 04:58 PM
i actually like all the calcs on GEDmatch each informative on its own, but for favorite my opinion is they are not on GEDmatch but on yourDNAportal maybe since i used everything one can do with V2K15 and K36 i like alot the K47 lukaszM and also the K33 MDLP would be cool if you could do as much with the K47 as with the K36;

Well 1 thing, the K47 test is one of the most accurate tests ive ever taken. And normally High K tests are super bad for me, like not even a exaggeration. This one tbh surprised me, it didn't give me too much European, it didn't underestimate my African/berber percentages, my Asian is just fine.

its a pretty good test ill give u that 100% XD

NarLFC
11-25-2018, 11:22 AM
Can you upload kits numbers to yourDNAportal?

Nino90
12-02-2018, 02:20 PM
Can you upload kits numbers to yourDNAportal?

You can just upload one Kit per account. I heard that they should add oracles too. But it gonna take some time.