What's a good population distance on calculators?

[Longbowman]

When on GEDmatch or similar, what's a 'good' distance between populations? IE, when you go on MDLP, they'll give you a list of the populations most similar to your genetics, like this:

Population (source) Distance
1 CITAL (Central-Italian) 4.05
2 SIC (Sicilian) 8.03
3 ASHK (Ashkenazi) 8.42
4 GRK (Greek) 11.69
5 NITAL (North-Italian) 12.73
6 CRS (Corsican) 18.1
7 CPR (Cypriot) 20.51
8 GGZ (Gagauz) 23.39
9 RMN (Romanian) 23.76
10 BLG (Bulgarian) 23.79
11 TRK (Turk) 25.85
12 MCD (Macedonian) 26.74
13 PRT (Portuguese) 27.33
14 MNT (Montenegrin) 29.29
15 SPN (Spaniard) 30.05
16 IBR (Iberian) 30.09
17 SRB (Serbian) 32.53
18 FRN (French) 33.43
19 ARM (Armenian) 34.68
20 BSN (Bosnian) 36.99

As you can see, the number to the right is the distance between your results and the population in question. Similar numbers are given for the mixed population results.

My question is: how low does a number have to be for you to consider yourself pure or nearly 100% from x-ethnicity? The lowest I've ever got (for a mixed population) was 0.5 (Central Italian with 1/8 Saami, who knew. That's being mixed ethnic for you).

What do you people think?

[Hevo]

MDLP results are skewed by calculator effect. Use Eurogenes.

What about Harappaworld?

[Longbowman]

Thanks for the information - I do use eurogenes - but MDLP was just an example. What I'm asking here is, what kind of plate distance would be expected for someone who pertains, say, 80-95% to that particular ethnicity? My question would be the same on Eurogenes.

[Gospodine]

You are quite Southern-shifted for a... figurative Brit, I suppose. No offense. Which part of your ancestry is the dominant side?

[Longbowman]

What makes you think this is mine? ;)

No offence taken - unless you have something against Romans? None of my ancestry forms a simply majority. I guess Italy is in the middle of England and Ukraine and the Ashkenazi/Sephardic urheimat etc.

What scores do you get for closest single population distances on most calculators?

[Gospodine]

What makes you think this is mine?

The results to me, fit someone of your N/E/S/W admixture, quite well.


What scores do you get for closest single population distances on most calculators?

I've yet to run any calculators actually; I'm still waiting on McDonald's BGA analysis to come back and GEDMatch is backlogged to high hell. I'll post them all in one hit when I get them. My 23andme results are up though. (http://www.theapricity.com/forum/showthread.php?102735-Gospodine-s-23andme-results)

[Longbowman]

Then you're NO USE TO ME, peasant!

When you get them, post them. I really want to know what a close distance is. Usually my closest single population hit is about 6 and to Central Italians. Then I get higher-scoring mixed results with English and AJs (unsurprising) and even closer ones with Italians and Arctic populations of some kind. I'm wondering if I should take that as signifying I do have some kind of Arctic ancestry or not, or if I should just dismiss it all.

[Longbowman]

The results to me, fit someone of your N/E/S/W admixture, quite well.

I've yet to run any calculators actually; I'm still waiting on McDonald's BGA analysis to come back and GEDMatch is backlogged to high hell. I'll post them all in one hit when I get them. My 23andme results are up though. (http://www.theapricity.com/forum/showthread.php?102735-Gospodine-s-23andme-results)

You know GEDmatch is back up again, right?

We're neanderthal admixture twins! 2.7% also, although seemingly less than yours because I'm 49th percentile. I'm not sure how to take the global similarity map - I plot in Western Ukraine when I'd have thought the rest of my ancestry would have dragged me west and south.

[Graham]

I'm not sure exactly but I'll show you what a Brit(Myself) scores for the area I come from. To give perspective
K13 Eurogenes1 West_Scottish 3.3
2 Danish 4.01
3 Irish 4.07
4 Orcadian 4.12
5 Southeast_English 5.07
6 Southwest_English 5.08
7 Dutch 6.79
8 Norwegian 7.47
9 Swedish 8.99
10 German 9.32
K15 Eurogenes1 Orcadian 4.08
2 West_Scottish 6.03
3 Danish 6.43
4 Irish 7.22
5 Southeast_English 7.51
6 Norwegian 7.61
7 Dutch 7.66
8 Southwest_English 8.47
9 Swedish 9.59
10 German 11.7
Dodecad V31 Irish (Dodecad) 5.53
2 Cornwall (1000 Genomes) 8.28
3 British_Isles (Dodecad) 8.49
4 British (Dodecad) 8.53
5 Norwegian (Dodecad) 10.16
6 Kent (1000 Genomes) 10.45
7 Dutch (Dodecad) 12.27
8 Mixed_Germanic (Dodecad) 13.62
9 Swedish (Dodecad) 13.88
10 Orkney (1000 Genomes) 18.85MDLP World 221 Welsh (derived) 3.3
2 German_V (derived) 3.53
3 Norwegian_V (derived) 4.68
4 Austrian (derived) 4.71
5 British (derived) 5.1
6 Swedish (derived) 5.48
7 Orcadian (derived) 5.82
8 German-North (derived) 6.09
9 Hungarian (derived) 7.07
10 German (derived) 7.34HarappaWorld1 british (1000genomes) 1.26
2 orcadian (hgdp) 2.51
3 n-european (xing) 4.45
4 french (hgdp) 11.27
5 hungarian (behar) 13.95
6 slovenian (xing) 14.7
7 ukranian (yunusbayev) 18.23
8 belorussian (behar) 20.91
9 mordovian (yunusbayev) 22.71
10 spaniard (behar) 22.87

[Gospodine]

Then you're NO USE TO ME, peasant!

Hey... how did the Ashes go? I didn't catch it actually... :bounce:


You know GEDmatch is back up again, right?

Yes and my holidays have only just started.


We're neanderthal admixture twins! 2.7% also, although seemingly less than yours because I'm 49th percentile. I'm not sure how to take the global similarity map - I plot in Western Ukraine when I'd have thought the rest of my ancestry would have dragged me west and south.

I really don't know how to take it either; I have a few people added on 23andme and I can tell you that Serbs cluster almost dead center in between Eastern and Southern Euro; which is considerably further south than me (despite being a close population). I'm quite matchy with a half-Turkish, half-Irish user.

My ancestry composition has also changed lately, and quite drastically. Now 23andme calculates me as predominantly Balkan 40-63%, with far less fragmentation of the remaining European admixture (it's all non-specific apart from 15 - 26% Eastern Euro). It also doesn't assign the remaining 2% or so of Non-Euro admixture.

[Longbowman]

[QUOTE]Hey... how did the Ashes go? I didn't catch it actually... :bounce:


That, my friend, was a good pun.


Yes and my holidays have only just started.


I get a month of for Christmas :)


I really don't know how to take it either; I have a few people added on 23andme and I can tell you that Serbs cluster almost dead center in between Eastern and Southern Euro; which is considerably further south than me (despite being a close population). I'm quite matchy with a half-Turkish, half-Irish user.

My ancestry composition has also changed lately, and quite drastically. Now 23andme calculates me as predominantly Balkan 40-63%, with far less fragmentation of the remaining European admixture (it's all non-specific apart from 15 - 26% Eastern Euro). It also doesn't assign the remaining 2% or so of Non-Euro admixture.


My global position hasn't changed but I gained some MENA, lost some European (those are the only two I score) gained some North European, lost some AJ, lost some Southern European and gained some Eastern.

The problem with 23andme is it doesn't go into the detail other calculators do.

[Longbowman]

I'm not sure exactly but I'll show you what a Brit(Myself) scores for the area I come from. To give perspective
K13 Eurogenes1 West_Scottish 3.3
2 Danish 4.01
3 Irish 4.07
4 Orcadian 4.12
5 Southeast_English 5.07
6 Southwest_English 5.08
7 Dutch 6.79
8 Norwegian 7.47
9 Swedish 8.99
10 German 9.32
K15 Eurogenes1 Orcadian 4.08
2 West_Scottish 6.03
3 Danish 6.43
4 Irish 7.22
5 Southeast_English 7.51
6 Norwegian 7.61
7 Dutch 7.66
8 Southwest_English 8.47
9 Swedish 9.59
10 German 11.7
Dodecad V31 Irish (Dodecad) 5.53
2 Cornwall (1000 Genomes) 8.28
3 British_Isles (Dodecad) 8.49
4 British (Dodecad) 8.53
5 Norwegian (Dodecad) 10.16
6 Kent (1000 Genomes) 10.45
7 Dutch (Dodecad) 12.27
8 Mixed_Germanic (Dodecad) 13.62
9 Swedish (Dodecad) 13.88
10 Orkney (1000 Genomes) 18.85MDLP World 221 Welsh (derived) 3.3
2 German_V (derived) 3.53
3 Norwegian_V (derived) 4.68
4 Austrian (derived) 4.71
5 British (derived) 5.1
6 Swedish (derived) 5.48
7 Orcadian (derived) 5.82
8 German-North (derived) 6.09
9 Hungarian (derived) 7.07
10 German (derived) 7.34HarappaWorld1 british (1000genomes) 1.26
2 orcadian (hgdp) 2.51
3 n-european (xing) 4.45
4 french (hgdp) 11.27
5 hungarian (behar) 13.95
6 slovenian (xing) 14.7
7 ukranian (yunusbayev) 18.23
8 belorussian (behar) 20.91
9 mordovian (yunusbayev) 22.71
10 spaniard (behar) 22.87

Perhaps I'm adopted and was born to some Roman family with minor Arctic or Amerindian admixture. Do you have any known non-British ancestry?

[Graham]

None at all. My Great grans Shetland side, may have some Scandinavians. I haven't seen any yet.

[Longbowman]

None at all. My Great grans Shetland side, may have some Scandinavians. I haven't seen any yet.

Hm, perhaps I'll have a word with my parents. Who knows. 0.5 distance for Italians + Arctic. Would explain my tolerance for incredible cold ^^

[Longbowman]

Longbowman, look at my (http://www.theapricity.com/forum/showthread.php?103004-Post-your-updated-Eurogenes-K13-GEDmatch-results&p=2118455&viewfull=1#post2118455) EUrogenes K13 results. Then compare to the Jaxman's results below. He is Sicilian/Irish/Finnish mix, yet he is closer to Austrians than I am to my people even though I'm only little mixed with neighbouring Slavs.
Also, notice mixed mode pop sharing results of a regular oracle may produce weird results. My mother certainly does not have East African ancestry.

Maybe you just want to believe that ;)

Yes, I suppose an English-AJ+Ukranian and others mix might give central Italian.

Beyond that, perhaps there is an Arctic component thrown into the mix.

Interesting results.

[Graham]

Mixed mode don't take it literally. Draw a line between the two nations, & work out where to stop from percentages. That's how I do it.

[Longbowman]

That isn't impossible, I agree. However, that would be very unusual for a Lithuanian-Belarusian-Russian. We certainly don't have Amerindian ancestry either...

The Amerindian is probably Siberian or Arctic which is very likely with a Baltic population.

[Herr Abubu]

My ancestry composition has also changed lately, and quite drastically. Now 23andme calculates me as predominantly Balkan 40-63%, with far less fragmentation of the remaining European admixture (it's all non-specific apart from 15 - 26% Eastern Euro). It also doesn't assign the remaining 2% or so of Non-Euro admixture.

23andme have added more reference populations from the Balkans (Albanians and Yugoslavs). Before the reference populations were only Greeks, Romanians and Bulgarians. That's - understandably - why I had so much non-specific Southern European admix before.

[Ibericus]

I would say in these calculators a distance above 10/11 means superficial closeness, that you deviate much from the average of that population, and anything under 5.00 is closeness, that you could be from that population.

[Longbowman]

I would say in these calculators a distance above 10/11 means superficial closeness, that you deviate much from the average of that population, and anything under 5.00 is closeness, that you could be from that population.

That's essentially what I supposed, meaning I mirror Central Italian populations, and even more closely, Central Italians with some Arctic/Inuit admixture.

[Black Wolf]

Yes generally if you are an admixed person it seems that you will fall into an area that is right in between your parent populations which makes sense. I am 25% Italian (Calabria), 25% Finnish and 50% Irish/British by known ancestry and I usually end up being closest to some Central European population when I use the Oracles.

Eurogenes K13:


Single Population Sharing:

# Population (source) Distance
1 Austrian 5.14
2 Hungarian 6.12
3 German 8.35
4 Serbian 11.21
5 Dutch 11.58
6 French 12.64
7 Swedish 13.05
8 Southeast_English 13.35
9 North_Swedish 13.72
10 Norwegian 13.74
11 Danish 13.87
12 Orcadian 14.26
13 Polish 14.8
14 Romanian 14.84
15 Irish 15.33
16 Southwest_English 15.56
17 Ukrainian_West 16.04
18 West_Scottish 16.04
19 Ukrainian 17.15
20 Southwest_Finnish 17.71

Eurogenes EUtest V2 K15:


Single Population Sharing:

# Population (source) Distance
1 German 7.5
2 Dutch 9
3 North_Swedish 10.56
4 Norwegian 10.62
5 Danish 10.67
6 French 10.82
7 Southeast_English 10.88
8 Swedish 10.96
9 Southwest_English 11.02
10 Hungarian 11.17
11 Orcadian 11.64
12 Austrian 11.99
13 Irish 12.14
14 West_Scottish 12.19
15 Serbian 13.33
16 Southwest_Finnish 13.42
17 Finnish 15.46
18 Romanian 17.26
19 Spanish_Galicia 17.57
20 Spanish_Cataluna 18.31

Eurogenes EUtest:


Single Population Sharing:

# Population (source) Distance
1 AT 7.74
2 West_&_Central_German 9.28
3 Serbian 10.1
4 NL 10.48
5 HU 11.11
6 DK 11.87
7 North_Swedish 12.34
8 English 12.44
9 NO 12.52
10 South_&_Central_Swedish 12.65
11 Cornish 13.34
12 Orcadian 13.94
13 RO 14.27
14 FR 14.3
15 IE 14.63
16 Scottish 15.99
17 South_Finnish 17.13
18 PL 18.09
19 PT 18.79
20 North_Italian 19.06

Eurogenes Jtest:


Single Population Sharing:

# Population (source) Distance
1 AT 7.68
2 West_&_Central_German 9.43
3 Serbian 10.23
4 NL 10.57
5 HU 11.1
6 DK 11.83
7 NO 12.39
8 English 12.53
9 South_&_Central_Swedish 12.58
10 North_Swedish 12.74
11 Cornish 13.37
12 Orcadian 13.99
13 RO 14.18
14 FR 14.24
15 IE 14.69
16 Scottish 15.89
17 South_Finnish 16.86
18 PL 18.03
19 PT 18.62
20 North_Italian 19.04

[Argang]

https://docs.google.com/spreadsheet/ccc?key=0Aqr2nbGXpVFndGlMazRoUXptekRSR2VUc2dpdVM0e nc#gid=0

On an individual level, you see higher than average results of East Asian, Northeast African etc. in individuals from numerous different areas.

It's probably still noise, there's no true fixed line between 1% and 1,01% that makes former noise and latter "old admixture".

[Longbowman]

https://docs.google.com/spreadsheet/ccc?key=0Aqr2nbGXpVFndGlMazRoUXptekRSR2VUc2dpdVM0e nc#gid=0

On an individual level, you see higher than average results of East Asian, Northeast African etc. in individuals from numerous different areas.

It's probably still noise, there's no true fixed line between 1% and 1,01% that makes former noise and latter "old admixture".

In my case though it's a) well above 1% (but below 10%) and very, very constant. Ie, every mixed population iteration will be at least 50% some kind of Italian and <10% Arctic.

[Argang]

What calculator gives you those oracle results though? MDLP, Dodecad and Harappaworld are all calculator effected, and will thus not show reliable results for non-project members.

Does McDonald show any amerindian or far eastern/siberian fits for you?

[Longbowman]

What calculator gives you those oracle results though? MDLP, Dodecad and Harappaworld are all calculator effected, and will thus not show reliable results for non-project members.

Does McDonald show any amerindian or far eastern/siberian fits for you?

I've sent my file to McDonald but haven't got results back, yet. This calculator was MDLP but the other ones give similar results.

[Damião de Góis]

MDLP guesses me as Swiss, so i think i could say that it's the least accurate one. Dodecad usually guesses me as spaniard because they have Spain broken down into regions, and it considers some spanish regions closer to me than Portugal. Eurogenes also has Spain broken down now, but it still manages to guess me as Portuguese, like in K13:


1 Portuguese 2.47
2 Spanish_Extremadura 3.4
3 Spanish_Cataluna 3.71
4 Spanish_Murcia 3.82
5 Spanish_Castilla_Y_Leon 3.89
6 Spanish_Galicia 4.22
7 Spanish_Valencia 4.44
8 Spanish_Andalucia 4.94
9 Spanish_Castilla_La_Mancha 5.17
10 Spanish_Cantabria 6.44
11 Spanish_Aragon 7.94
12 North_Italian 8.53
13 Southwest_French 8.75
14 French 9.38
15 Tuscan 14.39
16 French_Basque 19.37
17 Southeast_English 19.82
18 German 19.84
19 Romanian 19.96
20 Serbian 20.02

[Longbowman]

MDLP guesses me as Swiss, so i think i could say that it's the least accurate one. Dodecad usually guesses me as spaniard because they have Spain broken down into regions, and it considers some spanish regions closer to me than Portugal. Eurogenes also has Spain broken down now, but it still manages to guess me as Portuguese, like in K13:

They all give me the same results, give or take a percentage point (maybe South Italian not Central, maybe Evenks or Sami or Greenlandic). Eurogenes is no different.

[Damião de Góis]

They all give me the same results, give or take a percentage point (maybe South Italian not Central, maybe Evenks or Sami or Greenlandic). Eurogenes is no different.

That's not very usual, from what i've seen they usually give different results. It could be because your ancestry is very diverse and these calculators don't work as well for those cases.

[Longbowman]

That's not very usual, from what i've seen they usually give different results. It could be because your ancestry is very diverse and these calculators don't work as well for those cases.

They give different results, but only very marginal ones - different parts of Italy, different parts of the far North. The ancestry percentages are more varied, but they nearly always have a large (Central) Italian percentage of over 50%. It must be the mixed ancestry thing.

[Argang]

Can you post your Eurogenes K13 result? That should be accurate enough to get an idea whether the "eastern" admixture is real or not.

[Longbowman]

Can you post your Eurogenes K13 result? That should be accurate enough to get an idea whether the "eastern" admixture is real or not.

35.17% EAST MED
19.58% WEST MED
17.56% NORTH ATLANTIC
12.37% WEST ASIAN
5.71% BALTIC
5.56% RED SEA
1.27% SIBERIAN
0.90% NORTHEAST AFRICAN
0.79% SUBSAHARAN AFRICAN
0.56% EAST ASIAN
0.52% AMERINDIAN

The K13 oracles say pretty much the same as the rest of them.

[Argang]

Ukrainian average siberian is 1,41% in that calculator, West Ukrainian average is 1,36%.

Your combined "eastern" (Oceanian + Amerindian + East Asian + Siberian) is lower than average for any Ukrainian area.

Based on that I'd rule out any abnormal arctic ancestry and say oracles are messing up. But who knows, maybe McDonald will disagree.;)

[Longbowman]

Ukrainian average siberian is 1,41% in that calculator, West Ukrainian average is 1,36%.

Your combined "eastern" (Oceanian + Amerindian + East Asian + Siberian) is lower than average for any Ukrainian area.

Based on that I'd rule out any abnormal arctic ancestry and say oracles are messing up. But who knows, maybe McDonald will disagree.;)

Sure but presumably the average Ukrainian actually does have some Arctic ancestry?

[Graham]

35.17% EAST MED
19.58% WEST MED
17.56% NORTH ATLANTIC
12.37% WEST ASIAN
5.71% BALTIC
5.56% RED SEA
1.27% SIBERIAN
0.90% NORTHEAST AFRICAN
0.79% SUBSAHARAN AFRICAN
0.56% EAST ASIAN
0.52% AMERINDIAN

The K13 oracles say pretty much the same as the rest of them.

Finns, North Swedes & Russians score Siberian and then Amerindian, East Asian noise. It'll come from your ukranian side.

Btw anything under 1% Just ignore it. Noise isn't creditable.

[Longbowman]

Finns, North Swedes & Russians score Siberian and then Amerindian, East Asian noise. It'll come from your ukranian side.

Btw anything under 1% Just ignore it. Noise isn't creditable.

Absolutely, but 1.27% isn't under 1% (I don't think I'm subsaharan or Amerindian or East Asian).

[Argang]

1,27% is close enough to 1%. As I wrote, there's no magic line there.

Ukrainians have some tiny siberian input, but to put things into perspective, Orcadian average "eastern" admixture is greater than yours.

[Graham]

Absolutely, but 1.27% isn't under 1% (I don't think I'm subsaharan or Amerindian or East Asian).

Siberian. It shouldn't be strange in the North East at least.

https://sites.google.com/site/grpadmscand/N-Distr-Rootsi.jpg

[Longbowman]

1,27% is close enough to 1%. As I wrote, there's no magic line there.

Ukrainians have some tiny siberian input, but to put things into perspective, Orcadian average "eastern" admixture is greater than yours.

Fair enough. But as you say, if we were to add the related far-eastern Amerindian and Siberian and East Asian admixture, that's 2.35% 'eastern.'

[Longbowman]

Siberian. It shouldn't be strange in the North East at least.

Indeed.

Where did you get the maps?

[Graham]

Woops I edited your post first instead of quote haha. here

http://www.familytreedna.com/public/N%20Y-DNA%20Project

[Longbowman]

Woops I edited your post first instead of quote haha. here

http://www.familytreedna.com/public/N%20Y-DNA%20Project

It happens ;)

Thanks!

[Argang]

Fair enough. But as you say, if we were to add the related far-eastern Amerindian and Siberian and East Asian admixture, that's 2.35% 'eastern.'

Indeed, but that's a number possible for 100% northwest european.

North Swedes, Estonians, Finns, North Russians and Erzya are all within 5-10% range.

[Harkonnen]

yes. But Polako told me East African % must be something ancient as well.

Obviously it is impossible for it to be real ancient admixture (in your mother), for if it were ancient it would have to show on every and each one with similar background. It would be on all of us. I guess it could be possible for it to be real ancient admixture on all of us, that for some reason shows only in your mother.

[Longbowman]

Indeed, but that's a number possible for 100% northwest european.

North Swedes, Estonians, Finns, North Russians and Erzya are all within 5-10% range.

Fair enough. Is there some chart I can see that compares average admixture by ethnicity?

[Argang]

Here. (https://docs.google.com/spreadsheet/ccc?key=0Aqr2nbGXpVFndG1ra0p4UkI0b0d0b25YWTFHM0lHa mc&usp=sharing#gid=0)

[Hevo]

I would say in these calculators a distance above 10/11 means superficial closeness, that you deviate much from the average of that population, and anything under 5.00 is closeness, that you could be from that population.

What if like 4 countries are under 5.00?

[Harkonnen]

Maybe go and ask Polako yourself why exactly it is ancient.

Why this rude tone? Maybe apologize me maybe.

[Oneeye]

Being a New Worlder of a mix of Northwestern Euro immigrants, here are the distances I score:

Eurogenes K13

1 German 3.57
2 Dutch 4.66
3 Southeast_English 4.8
4 Orcadian 5.84
5 Danish 6.07
6 Irish 6.37
7 Southwest_English 6.51
8 West_Scottish 7.06
9 Norwegian 8.46
10 Swedish 9.21



Eurogenes K15


1 Dutch 4.56
2 Orcadian 5.32
3 Norwegian 6.26
4 Danish 6.3
5 Southwest_English 6.48
6 Southeast_English 6.5
7 German 6.94
8 West_Scottish 7.18
9 Irish 7.59
10 Swedish 7.97



Dodecad V3


1 CEU (HapMap) 2.63
2 N._European (Xing) 4.05
3 Argyll (1000 Genomes) 4.38
4 Orcadian (HGDP) 4.5
5 Orkney (1000 Genomes) 4.87
6 German (Dodecad) 7.5
7 Mixed_Germanic (Dodecad) 9.2
8 French (HGDP) 9.34
9 French (Dodecad) 9.83
10 Dutch (Dodecad) 10.34


MDLP World 22


1 German_V (derived) 4.71
2 Austrian (derived) 5.28
3 CEU_V (derived) 5.29
4 CEU (derived) 5.39
5 Welsh (derived) 5.91
6 British (derived) 6.29
7 German-South (derived) 6.46
8 Norwegian_V (derived) 6.87
9 Hungarian (derived) 7.1
10 Orcadian (derived) 7.41



HarappaWorld


1 utahn-white (1000genomes) 4.16
2 n-european (xing) 5
3 utahn-white (hapmap) 5.34
4 british (1000genomes) 5.9
5 french (hgdp) 7.48
6 orcadian (hgdp) 8.3
7 hungarian (behar) 9.22
8 slovenian (xing) 10.69
9 ukranian (yunusbayev) 16.89
10 spaniard (behar) 19.68

[Atlantic Islander]

The closer the better, 5 at the max imo.

[Atlantic Islander]

MDLP guesses me as Swiss, so i think i could say that it's the least accurate one. Dodecad usually guesses me as spaniard because they have Spain broken down into regions, and it considers some spanish regions closer to me than Portugal. Eurogenes also has Spain broken down now, but it still manages to guess me as Portuguese, like in K13:

MDLP was okay for me (world22):

1 Portuguese (derived) 5.36

Mixed mode is a bit of a mess.

[Damião de Góis]

MDLP was okay for me (world22):

1 Portuguese (derived) 5.36

Mixed mode is a bit of a mess.

World22 is completely off for me, which is why i don't pay much attention to MDLP:


# Population (source) Distance
1 Swiss (derived) 4.51
2 Italian_North (derived) 4.85
3 Provancestralal (derived) 5.6
4 Portugese (derived) 8.45
5 French (derived) 9.41
6 Montenegrin (derived) 9.51
7 Italian-North (derived) 9.76
8 Iberian (derived) 9.93
9 Bulgarian (derived) 10.05
10 Romania (derived) 10.12
11 German-South (derived) 10.29
12 Macedonian (derived) 10.37
13 Spaniard (derived) 10.75
14 Corsican (derived) 11.26
15 Gagauz (derived) 11.93
16 Serbian (derived) 12.18
17 Kosovar (derived) 12.68
18 Ashkenazim_V (derived) 13.84
19 Greek_South (derived) 15.3
20 British (derived) 15.52

[Jackson]

I would say <5 is very close (as in you would be quite normal in that population), <10 is relatively close - not typical but you could be that ancestry + something else, and greater than 15 is pretty much not of that population.
Although, some populations vary more than others, so this might not apply to areas that have a lot of internal variation, or who are quite far from their next nearest neighbors. Although i'd say that isn't neccesarily reliable in cases where you are quite mixed, or even a little. Obviously i'm not German but if you mix an English person with a little bit of something from southern Europe you get a (i'd guess western) German, although it picks this mixture up in the mixed mode, and usually places my normal population as Southeast English or Dutch which is more accurate.

Also it depends a lot on the samples too, if a population has very few samples, and it's quite a varied place, it might be less accurate for someone who's ancestry is not well represented. Like if you were from southern France and there were only samples from northern France, you'd probably come out as French + Iberian or French + Italian, even though you were 100% French, but from a different area. I guess that's where the components come in handy, to see if you have any admixture that's very atypical in your population, or whether you just have an excess of one particular component that makes you atypical.

[RighNick]

Sorry to revive this topic after so long, but I find the question very interesting. I noticed that I am very far from most groups in single population mode and even in mixed mode my "distance" from my best match is around 7. Like longbowman, I was wondering what a typical distance was for someone solely of one ethnicity.

[Musso]

For mixed individuals such population distances aren't very useful (especially single population distances). Basically it gives you the population that is in the middle or average of your ancestries. That doesn't mean you are related to those populations per se.

If a person for example is half ethnicity x, half ethnicity y, they should look at the two or three population distance. Single population distance just isn't useful. A half Italian, half English person might be most closest to French in terms of population distance, but doesn't mean the person has French blood.

[RighNick]

For mixed individuals such population distances aren't very useful (especially single population distances). Basically it gives you the population that is in the middle or average of your ancestries. That doesn't mean you are related to those populations per se.

If a person for example is half ethnicity x, half ethnicity y, they should look at the two or three population distance. Single population distance just isn't useful. A half Italian, half English person might be most closest to French in terms of population distance, but doesn't mean the person has French blood.

Interesting, I actually have a situation very similar to the one you describe. I'm primarily a mix of Northern and Southern European, and I almost always come up as French despite only having limited French ancestry. You're certainly right about the two population mode being more accurate, at least in my case. In 2 population mode on eurogenes K13 I get "50% Norwegian +50% Spanish_Murcia @ 9.925". That's fairly close to my actual ancestry, but what of the distances in two or three population mode? The distance is much farther than other peoples I've seen. 9.925 is my closest distance while 4.092 is my mother's farthest. Why would that be the case? Am I just an outlier?

[nittionia]

My dad's distances on Eurogenes K13 and K15 are too much
https://i.imgur.com/iWDGUrY.png

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