PDA

View Full Version : What makes Eurogenes K36 so popular?



Apex n Harmony
10-18-2017, 04:34 PM
From what I've seen, Eurogenes K36 seems to be the most popular and trusted calculator (by far) when it comes to using fun tools like determining where you fall on a world map, determining your similarity to ancient populations, etc. Why is that? Are we not putting all our faith in the alleged accuracy of Eurogenes K36 results (and if we are, why do we trust the results so much)?

Apex n Harmony
10-20-2017, 04:54 PM
Giving this a little "bump" only because I imagine many people here have been wondering the same thing.

Tolan
10-21-2017, 03:39 AM
Even though its author does not seem to like it very much, the components are remarkable even if their names were very badly chosen, because they correspond to ancient components than our modern countries.

New calculators, created from ancient genomes, are not very reliable for modern populations. (This is my opinion),.
For example, English will be close to the Czechs.
Their components are too old to determine current populations

The second reason is the large number of averages per country and region collected by LukaszM.
No other calculator has so much average!
A good calculator with a large average number around the world! What more could we ask for?

MitchellSince1893
10-21-2017, 05:12 AM
From what I've seen, Eurogenes K36 seems to be the most popular and trusted calculator (by far) when it comes to using fun tools like determining where you fall on a world map, determining your similarity to ancient populations, etc. Why is that? Are we not putting all our faith in the alleged accuracy of Eurogenes K36 results (and if we are, why do we trust the results so much)?

For me it compares favorably (once the values are plugged into Tolan's tool http://gen3553.pagesperso-orange.fr/ADN/similitude.htm) with my family's known genealogical records....mostly British with mother having more Scandinavian and German Ancestors than my father.

From left to right, father, me, mother.
https://i.pinimg.com/originals/a1/c5/bc/a1c5bc8535344a560eb6500c89bd840f.png

Plus K36 results and Tolan's tool let's you create a map.

anglesqueville
10-21-2017, 08:34 AM
Let's call "good variable" any function (x,y) -----> z whose spatial representation is a more or less deformed half hyperboloid, something like that:
19393

The "nature" of a good variable does not matter. It can even be absolutely non genetical: if it can be represented in that way ( roughly: a decreasing gradient from a single center), it is a good variable. The statistics of the spaghettis consumption provide likely a good variable, with a center on Italy. Take N good variables, N large enough. 36 is obviously a good N value. You get a good "calculator", and you can start drawing magnificent maps.

Generalissimo
10-21-2017, 11:34 AM
Even though its author does not seem to like it very much, the components are remarkable even if their names were very badly chosen, because they correspond to ancient components than our modern countries.

Au contraire, I like it very much. And the names are just, well...names. Feel free to change them.


New calculators, created from ancient genomes, are not very reliable for modern populations. (This is my opinion),.
For example, English will be close to the Czechs.
Their components are too old to determine current populations

Indeed, because when they're properly designed, they precede modern day drift that today separates, say, English from Czechs. Do you see now?


The second reason is the large number of averages per country and region collected by LukaszM.
No other calculator has so much average!
A good calculator with a large average number around the world! What more could we ask for?

LukaszM has done an awesome job. That's probably because he's Polish, and Polish people are awesome. And I have already rewarded LukaszM indirectly for his efforts.

tooltime
10-29-2017, 01:02 PM
I like the K36 calculator because of the many regional breakdowns. My results have usually been "boring" in most calculators (typical Russian results with lots of North Atlantic and Baltic and some general Mediterranean/Asian). Here's my K13:
Population
North_Atlantic 23.94
Baltic 45.86
West_Med 10.66
West_Asian 4.01
East_Med 9.12
Red_Sea -
South_Asian -
East_Asian -
Siberian 4.83
Amerindian 1.13
Oceanian -
Northeast_African 0.34
Sub-Saharan 0.10

K36 hints at more specific breakdowns of ancient Southern Europe and Mediterranean regions:
Population
Amerindian 0.12
Arabian -
Armenian -
Basque -
Central_African -
Central_Euro 6.68
East_African -
East_Asian -
East_Balkan 10.14
East_Central_Asian -
East_Central_Euro 22.12
East_Med 0.60
Eastern_Euro 23.23
Fennoscandian 9.62
French 1.97
Iberian 1.40
Indo-Chinese -
Italian 1.82
Malayan -
Near_Eastern 2.51
North_African -
North_Atlantic 6.34
North_Caucasian 0.34
North_Sea 8.41
Northeast_African -
Oceanian -
Omotic -
Pygmy -
Siberian 0.81
South_Asian -
South_Central_Asian -
South_Chinese -
Volga-Ural 3.68
West_African -
West_Caucasian -
West_Med 0.24

Apex n Harmony
10-29-2017, 04:30 PM
I like the K36 calculator because of the many regional breakdowns. My results have usually been "boring" in most calculators (typical Russian results with lots of North Atlantic and Baltic and some general Mediterranean/Asian). Here's my K13:
Population
North_Atlantic 23.94
Baltic 45.86
West_Med 10.66
West_Asian 4.01
East_Med 9.12
Red_Sea -
South_Asian -
East_Asian -
Siberian 4.83
Amerindian 1.13
Oceanian -
Northeast_African 0.34
Sub-Saharan 0.10

K36 hints at more specific breakdowns of ancient Southern Europe and Mediterranean regions:
Population
Amerindian 0.12
Arabian -
Armenian -
Basque -
Central_African -
Central_Euro 6.68
East_African -
East_Asian -
East_Balkan 10.14
East_Central_Asian -
East_Central_Euro 22.12
East_Med 0.60
Eastern_Euro 23.23
Fennoscandian 9.62
French 1.97
Iberian 1.40
Indo-Chinese -
Italian 1.82
Malayan -
Near_Eastern 2.51
North_African -
North_Atlantic 6.34
North_Caucasian 0.34
North_Sea 8.41
Northeast_African -
Oceanian -
Omotic -
Pygmy -
Siberian 0.81
South_Asian -
South_Central_Asian -
South_Chinese -
Volga-Ural 3.68
West_African -
West_Caucasian -
West_Med 0.24

That may be true, but the more specific and detailed results are, the more inaccurate they're likely to be (or at least that's what I've always read).

Apex n Harmony
10-29-2017, 04:46 PM
Au contraire, I like it very much. And the names are just, well...names. Feel free to change them.



Indeed, because when they're properly designed, they precede modern day drift that today separates, say, English from Czechs. Do you see now?



LukaszM has done an awesome job. That's probably because he's Polish, and Polish people are awesome. And I have already rewarded LukaszM indirectly for his efforts.

One thing I've always found a bit strange is how most of the components are region-specific, while a select few are "country" or "nation" specific (e.g., Armenian, Italian, French). I would imagine this is because those specific components are just that unique. Is that right? For example, given that there are already components for West Caucasian and North Caucasian, can we deduce that the Armenian component is a standalone because the original Armenian community was that much more genetically unique than neighboring Caucasian communities? (I have an Armenian friend who scores 14.68% Armenian, 18.44% North Caucasian, and 11.09% West Caucasian. Does that mean that, per this calculator, she matches an "original" Armenian about 14.68%?)

vettor
10-29-2017, 05:25 PM
For me it compares favorably (once the values are plugged into Tolan's tool http://gen3553.pagesperso-orange.fr/ADN/similitude.htm) with my family's known genealogical records....mostly British with mother having more Scandinavian and German Ancestors than my father.

From left to right, father, me, mother.
https://i.pinimg.com/originals/a1/c5/bc/a1c5bc8535344a560eb6500c89bd840f.png

Plus K36 results and Tolan's tool let's you create a map.

how did you get the map shading

and should one remove ( as per right site of the tool ) the areas where ones results bears a zero number ?..........example I get no north-africa results in k36, should I remove north african from the options

Edit; ....I have North-Italy as the only red square ...........if I remove north-africa from the data because I have no north-africa , my red square shifts to south-germany only

rwtodd
11-20-2017, 05:09 PM
Your mother's map corresponds almost perfectly with mine. I will echo another poster's question: How did you make these very cool heat maps?

Dibran
12-20-2017, 04:51 PM
Your mother's map corresponds almost perfectly with mine. I will echo another poster's question: How did you make these very cool heat maps?

Agreed. Also very curious. Tried with paint but it looks crappy.

JerryS.
12-24-2017, 05:33 AM
categorizing English as very similar to Czech?

Exosuits
12-30-2017, 01:14 PM
With the high population components, I find it quite useful for segment runs with the byseg or bychr utility tools.

Randwulf started a very interesting thread about it

https://anthrogenica.com/showthread.php?7557-Attempts-To-Use-Euro-K36-As-Purposed&p=165530&viewfull=1#post165530

Kurki
01-25-2018, 01:18 AM
Even though its author does not seem to like it very much, the components are remarkable even if their names were very badly chosen, because they correspond to ancient components than our modern countries.

New calculators, created from ancient genomes, are not very reliable for modern populations. (This is my opinion),.
For example, English will be close to the Czechs.
Their components are too old to determine current populations

The second reason is the large number of averages per country and region collected by LukaszM.
No other calculator has so much average!
A good calculator with a large average number around the world! What more could we ask for?

Is this using ancient genomes? This one made me quite Iberian and French, far more than most, though it shows up once in a while on GEDmatch. I am not French, but Scandinavian. But there is overlap with Viking DNA. This happens with Orkneys as well. More modern migrations from Scotland to Norway are not super uncommon, but I have wondered about that French and Iberian connection as probably possibly being Viking-esque and not even common DNA in present day Norway.

I think the specificity is interesting, yet vague... pinpointing French, but then overlapping the whole northwest a few ways, so I'm not sure where Denmark was covered. My family is quite Danish and I don't know... I assume that was under Fennoscandia. But there is a North Sea and a North Atlantic group on here. Both also high. But that is probably British Isles and then Germany/France. All Viking stuff maybe. But those attempts at specifics areas end up overlapped when you try to look up admixture maps for places like Denmark...which I know is Scandinavian, but I also know is dropped from some Scandinavian reference populations (like it just doesn't exist), so I don't know... maybe it's "North Sea".. no big deal really, but I can't find a North Sea map. I assume North Sea is mostly specific to British Isles when North Atlantic and Fennoscandia is also added in. ?

RobinBMc
01-25-2018, 06:58 PM
I don't put my trust in any ethnicity report, but the more specific it is/the more break down, the bigger grain of salt I take it with. K36 is definitely highly speculative, if you ask me. I get so many small amounts in different places, I can't take it very seriously.

Amerindian -
Arabian -
Armenian -
Basque 1.11
Central_African -
Central_Euro 5.87
East_African -
East_Asian -
East_Balkan 5.13
East_Central_Asian -
East_Central_Euro 1.11
East_Med 4.95
Eastern_Euro -
Fennoscandian 8.27
French 6.86
Iberian 8.49
Indo-Chinese -
Italian 16.68
Malayan -
Near_Eastern 4.41
North_African -
North_Atlantic 14.50
North_Caucasian 2.87
North_Sea 12.69
Northeast_African -
Oceanian -
Omotic -
Pygmy -
Siberian -
South_Asian -
South_Central_Asian -
South_Chinese -
Volga-Ural -
West_African -
West_Caucasian 3.76
West_Med 3.29

yelmex
01-27-2018, 06:04 PM
How can I use population sharing for genetic distance from this test?

Tolan
01-27-2018, 07:01 PM
I don't put my trust in any ethnicity report, but the more specific it is/the more break down, the bigger grain of salt I take it with. K36 is definitely highly speculative, if you ask me. I get so many small amounts in different places, I can't take it very seriously.



Yes, everyone should know it!
And this is valid for all calculators on gedmatch.

They don't give your composition from the populations of today, but from ancients populations. Hence my remark on the chosen names of the components that can be misleading.
Only the comparison with the country averages makes it possible to know from which average we are getting closer.

The results of the DNA companies (23andMe, FTDNA, AncestryDna ect ..) have for ambition to give you directly your ancestral composition.
It's different from the gedmatch calculators, DNALand and others

Vestri
02-03-2018, 10:13 PM
For me it compares favorably (once the values are plugged into Tolan's tool http://gen3553.pagesperso-orange.fr/ADN/similitude.htm) with my family's known genealogical records....mostly British with mother having more Scandinavian and German Ancestors than my father.

From left to right, father, me, mother.
https://i.pinimg.com/originals/a1/c5/bc/a1c5bc8535344a560eb6500c89bd840f.png

Plus K36 results and Tolan's tool let's you create a map.

Mitchell may I ask, how were you able to colour your maps? That looks briliant and I could do with that myself.

Thanks :).

Hayden
02-05-2018, 05:34 AM
Mitchell may I ask, how were you able to colour your maps? That looks briliant and I could do with that myself.

Thanks :).

They do look very nice Mitchell. I made a mask in illustrator so I can make something similar quickly if anyone wants. Just need the k36 numbers so I can use Tolans beautiful map tool :beerchug:

I just made one for my grandmother's
https://farm5.staticflickr.com/4711/26217731908_60b93b39a1_t.jpg (https://farm5.staticflickr.com/4711/26217731908_5ab3955828_o.png)

alexfritz
02-05-2018, 07:48 AM
They do look very nice Mitchell. I made a mask in illustrator so I can make something similar quickly if anyone wants. Just need the k36 numbers so I can use Tolans beautiful map tool :beerchug:

I just made one for my grandmother's
https://farm5.staticflickr.com/4711/26217731908_60b93b39a1_t.jpg (https://farm5.staticflickr.com/4711/26217731908_5ab3955828_o.png)

those maps look awesome, i would like one very much;
if it is not much trouble thanks alot

Population
Amerindian -
Arabian -
Armenian 2.05
Basque 2.99
Central_African -
Central_Euro 6.22
East_African -
East_Asian -
East_Balkan 1.80
East_Central_Asian -
East_Central_Euro -
East_Med 7.21
Eastern_Euro 5.61
Fennoscandian 5.81
French 3.47
Iberian 13.00
Indo-Chinese -
Italian 25.25
Malayan -
Near_Eastern -
North_African -
North_Atlantic 8.13
North_Caucasian -
North_Sea 9.72
Northeast_African -
Oceanian -
Omotic -
Pygmy -
Siberian -
South_Asian -
South_Central_Asian -
South_Chinese -
Volga-Ural -
West_African -
West_Caucasian 1.24
West_Med 7.50

Hayden
02-05-2018, 10:17 AM
those maps look awesome, i would like one very much;
if it is not much trouble thanks alot

Happy to make use of the template I made, hopefully I didn't transpose anything.
https://farm5.staticflickr.com/4655/25223254927_8dfd9cf27d_t.jpg (https://farm5.staticflickr.com/4655/25223254927_d495b0341c_o.png)

alexfritz
02-05-2018, 11:00 AM
Happy to make use of the template I made, hopefully I didn't transpose anything.
https://farm5.staticflickr.com/4655/25223254927_8dfd9cf27d_t.jpg (https://farm5.staticflickr.com/4655/25223254927_d495b0341c_o.png)

spot on, thank you very much

Sebbo
04-23-2018, 08:56 AM
Is Eurogenes K36 considered better than other Eurogenes calculators?

ianz91
04-28-2018, 06:00 AM
Happy to make use of the template I made, hopefully I didn't transpose anything.
https://farm5.staticflickr.com/4655/25223254927_8dfd9cf27d_t.jpg (https://farm5.staticflickr.com/4655/25223254927_d495b0341c_o.png)


Can you please please do mine?

Basque 1.80
Central Euro 6.74
East Balkan 5.42
East Central Euro 7.77
East Med 1.12
Eastern Euro 4.93
Fennoscandian 9.67
French 7.67
Iberian 19.67
Italian 8.98
North Atlantic 13.35
North Sea 11.50
West Med 1.36

Vashti
05-03-2018, 04:13 PM
That is a beautiful map! I willhave to now run my newest K36 :)

Askye
05-21-2018, 10:59 AM
K36 is the most off and incorrect of any calculator out there for me. I find it sad when I see how people rely on it so much. Grain of salt.

cvolt
05-21-2018, 11:05 AM
My highest percentage on k36 is Iberian at ~20% lol

ianz91
05-22-2018, 12:33 PM
K36 is the most off and incorrect of any calculator out there for me. I find it sad when I see how people rely on it so much. Grain of salt.


Honestly, I take most GEDMATCH calculators with a grain of salt, I do like the oracle results they give though.

digital_noise
05-22-2018, 05:12 PM
I might be misunderstanding it but I do not think it is supposed to be used like you would use the K13 or K15 calculator. It seems to function well if used like a plot point for data for an "unofficial" oracle.

Xtian
09-21-2018, 09:12 AM
Honestly, I take most GEDMATCH calculators with a grain of salt, I do like the oracle results they give though.

Gedmatch is best used when comparing yourself to the spreadsheet to certain populations.

I wish the commercial DNA companies did something similar. They should allow to you to compare yourself to a blended average of people from different ethnicities.

Xtian
09-21-2018, 09:13 AM
duplicate

Teutorigos
12-24-2018, 11:12 PM
From what I've seen, Eurogenes K36 seems to be the most popular and trusted calculator (by far) when it comes to using fun tools like determining where you fall on a world map, determining your similarity to ancient populations, etc. Why is that? Are we not putting all our faith in the alleged accuracy of Eurogenes K36 results (and if we are, why do we trust the results so much)?

First of all Americans take things too lightly it is like they are not mature or adult-like mentally. This is not 'fun stuff' this is serious science.

Eurogenes k36 is complete inferior crap it is too ambitious and complicated :

"Everything should be made as simple as possible but not simpler." -- Albert Einstein


Having 36 components is the opposite of simple ; also , it has my north sea component around 20 , which is absurd , meanwhile k15 has my North Sea at 35.89 which is much more reasonable since I am Southern British.

All GEDmatch calculators are inferior crap compared to PuntDNAL with one caveat : Davidski seems to be on to something with the so called 'calculator effect' but since I am British, hence sophisticated , it does not affect me on PUNTDNAL because I know how to ignore extraneous information. Anything 5 and below you can pass as a native of that population. As far as the Oceanian thing I view that as the British equivalent of American native American admixture it is from the British empire days and is not particular significant, mathematically, in my case :

puntDNAL K12 Ancient Oracle

Admix Results (sorted):

# Population Percent
1 European_HG 42.24
2 Anatolian_NF 34.06
3 Caucasus_HG 18.94
4 Near_East 1.75
5 Oceanian 1.58


Finished reading population data. 55 populations found.
12 components mode.

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

Least-squares method.

Using 1 population approximation:
1 Unetice_EBA_I0117 @ 3.558241
2 Halberstadt_LBA_I0099 @ 3.891102
3 Nordic_LN_SG_RISE97 @ 4.618601
4 Alberstedt_LN_I0118 @ 4.726390

Unetice Culture early Bronze age around the region which corresponds to Esperstedt Saxony-Anholt Germany. Halberstadt Saxony Anhalt Germany late bronze age .


puntDNAL K12 Modern Oracle results:
puntDNAL K12 Modern Oracle

Kit A703943

Admix Results (sorted):

# Population Percent
1 European_HG 42.24
2 Anatolian_NF 34.06
3 Caucasus_HG 18.94
4 Near_East 1.75
5 Oceanian 1.58
6 South_Asian 0.99
7 South_African_HG 0.24
8 Beringian 0.21

Single Population Sharing:

# Population (source) Distance
1 English_South 2.54

https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcQt1reBab7rkZ7ycL1nZP5K8EoYyPJs6 Ztm8iXUajLaCHWbM7_dSg


https://www.youtube.com/watch?v=926fiS2kvt8

JoeyP37
12-25-2018, 01:41 AM
I am an American whose ethnic background is as follows-3/8 English, 3/16 Irish, 7/32 German, 1/8 Sicilian, 1/16 French, 1/16 Dutch, and 1/32 Welsh + Native American. Part of my Sicilian is Arbereshe/Greek since I get Balkan at 23andMe. I did the puntDNAL global K13 and got this:
1 NE_Europe 44.59
2 SW_Europe 38.16
3 West_Asia 11.1
4 South_Asia 2.58
5 Siberia 1.42
6 SW_Asia 1.18
7 Americas 0.64
8 West_Africa 0.33

Single Population Sharing:

# Population (source) Distance
1 German_South 3.48
2 Belgian 4.92
3 English 5.16
4 Scottish 5.34
5 Orcadian 5.9

Teutorigos
12-25-2018, 08:59 AM
I am an American whose ethnic background is as follows-3/8 English, 3/16 Irish, 7/32 German, 1/8 Sicilian, 1/16 French, 1/16 Dutch, and 1/32 Welsh + Native American. Part of my Sicilian is Arbereshe/Greek since I get Balkan at 23andMe. I did the puntDNAL global K13 and got this:
1 NE_Europe 44.59
2 SW_Europe 38.16
3 West_Asia 11.1
4 South_Asia 2.58
5 Siberia 1.42
6 SW_Asia 1.18
7 Americas 0.64
8 West_Africa 0.33

Single Population Sharing:

# Population (source) Distance
1 German_South 3.48
2 Belgian 4.92
3 English 5.16
4 Scottish 5.34
5 Orcadian 5.9


Honestly, all GEDmatch calculator suite sets leave something to be desired but what do you expect for free ? When I say PUNTDNAL is superior I meant in context of using Eurogenes k36 for similarity to ancient populations and for so called 'mapping'. I have no opinion on specialized niches like Hrappaworld, Ethiohelix etc... As I have said , before, PuntDNAL k12 modern agrees with my interpretome mapping (at least it counts where it is the most important in pegging my single population dead on) etc... and as far as the the global PuntDNAL calculator it may not be the best in the suite, not sure, but it is a waste of my time since I am 75% British isles and 25% Alsace-Lorraine (Elssaserdeutsch+Frankish) and as I have already said southern England is no more than 25% kraut.