Modified K13 run for 4 Magyars and 2 Croats

[Jana]

I did manual conversion of their K13 results to 23andme V4 version, because they all tested with myHeritage or 23andme V5 chip which have snps cut in half.
Technique was the following: I compared my own 23andme V4 and Myheritage values, than applied exact value differences +/- to their results in same fashion like my results differ.
So, this should be close to how would their results looked like with more than 2x snps tested.

My Step Father from western Croatia


Stearsolina_StepDad,31.77,32.47,13.67,6.47,10.2,3. 31,0.17,0,1.87,0,0.4,0,0.11

https://i.ibb.co/1Q9MrT9/K13a.png (https://imgbb.com/)
https://i.ibb.co/jHY0TMs/Mk13.png (https://ibb.co/1bNyLqj)

My family friend from Eastern Croatia (she's probably 1/4 Czech, Slovak or Rusyn)


Manda,29.46,37,13.24,9.28,7.66,2.46,0,0,0.98,0,0.0 5,0,0

https://i.ibb.co/znwGBZ1/K13b.png (https://imgbb.com/)
https://i.ibb.co/4YZRTDC/mk13b.png (https://ibb.co/5nWh8zC)

My Hungarian friend from Budapest (she has Polish, Slovak and German ancestry) - in her myheritage run her German ancestry was not visible, but now it is


Blanka,34.94,31.03,13.75,3.49,11.58,3.23,0,0,1.3,0 .23,0,0,0.39

https://i.ibb.co/3kqgdYH/K13c.png (https://imgbb.com/)

https://i.ibb.co/M5D3Xm4/mk13c.png (https://imgbb.com/)

Her Hungarian husband from South-eastern Hungary, but I found out he has some ancestry from southern Transdanubia as well. His one parent was adopted so he doesn't know much about his origin except that he's Hungarian with some Slovak ancestry.
He's very German shifted and that could be explained by ancestry from Baranya, which was called Schwabische Turkei, so he's probably part Danube Swabian.


Blanka_Husband,35.78,30.31,15.55,7.69,8.6,0.13,0,0 ,1.36,0,0,0.19,0.5

https://i.ibb.co/RTT3tNF/K13d.png (https://imgbb.com/)

https://i.ibb.co/Qj76ZNY/mk13d.png (https://imgbb.com/)

TA user Universe (from eastern Hungary, he has recent Slovak ancestry)


Universe,30.68,36.45,11.38,4.71,12.83,2.31,0,0,1.1 5,0,0.23,0.72,0

https://i.ibb.co/0XZqwHJ/K13e.png (https://imgbb.com/)

https://i.ibb.co/NjdcCNx/mk13e.png (https://imgbb.com/)

TA user Pajkosbalna (I think she's from NE Hungary and has some Rusyn ancestry but not 100% sure)


Pajkosbalna,23.23,38.81,16.64,7.42,8.93,0.88,1.6,0 ,1.15,1.3,0,0,0

https://i.ibb.co/QD9G8R9/K13f.png (https://imgbb.com/)

https://i.ibb.co/MZv2tnK/mk13f.png (https://imgbb.com/)

Both Universe and Pajkosbalna had outlying Hungarian results with off the charts Baltic (and in case of Pajkos super low North Atlantic for Hungary), now their results look lot more balanced.
They remain one of most Slavic/Baltic shifted Hungarians ofcourse.

Conclusion is 23andme V5 and Myheritage raw data shiftes everyone results to the east because of low snp count, so I tried to fix that.

[Lucas]

Interesting, so you added/distracted to your father k13 components the same values as you have different?

[Jana]

Interesting, so you added/distracted to your father k13 components the same values as you have different?

Yes.

I did the same for Roy, his results make lot more sense now. He plotted like East Slav with V5 data despite he has no ancestry from Kresy.


Hello Roy,

I manually converted your K13 results to values you would probably get with 23andme V4 chip (which tested more than 2x more snps than V5), and to me these results look lot more logical.


Roy,28.81,44.27,12.08,5.45,3.1,1.52,0.79,0,0.67,2. 07,0,0.43,0.73

https://i.ibb.co/f4FxHLN/roy.png (https://imgbb.com/)

https://i.ibb.co/Yf6X3J8/roy2.png (https://imgbb.com/)

^^^^with my method applied his results look typically Polish now.

[Lucas]

Yes.

I did the same for Roy, his results make lot more sense now. He plotted like East Slav with V5 data despite he has no ancestry from Kresy.



^^^^with my method applied his results look typically Polish now.

But in his case you added/distracted your values too? Maybe post them I can try with some people.

[vbnetkhio]

I did manual conversion of their K13 results to 23andme V4 version, because they all tested with myHeritage or 23andme V5 chip which have snps cut in half.
Technique was the following: I compared my own 23andme V4 and Myheritage values, than applied exact value differences +/- to their results in same fashion like my results differ.
So, this should be close to how would their results looked like with more than 2x snps tested.

My Step Father from western Croatia


Stearsolina_StepDad,31.77,32.47,13.67,6.47,10.2,3. 31,0.17,0,1.87,0,0.4,0,0.11


looks potentially very useful, but it would be good if you collected more new/old chip pairs and checked if there's really a bias towards Baltic.
Hungarian individuals really can be high Baltic, their oracles will look nothing like an average Hhungarian's, but somebody always has to be on the edge of the cluster.
User Michal is also an east shifted Pole, does he have any ancestry from Kresy? Maybe Baltic ancestry from northern Poland could cause such a shift as well.

[Lucas]

looks potentially very useful, but it would be good if you collected more new/old chip pairs and checked if there's really a bias towards Baltic.
Hungarian individuals really can be high Baltic, their oracles will look nothing like an average Hhungarian's, but somebody always has to be on the edge of the cluster.
User Michal is also an east shifted Pole, does he have any ancestry from Kresy? Maybe Baltic ancestry from northern Poland could cause such a shift as well.

Michal3232 is mainly from Malopolska, but he has also older chip I think.

[vbnetkhio]

Michal3232 is mainly from Malopolska, but he has also older chip I think.

Stearsolina thinks the new (v5) chip gives inflated baltic, not the old one :)

[Leto]

I've seen quite a few Russian results with Genotek data that scored close to 30% North Atlantic which would mean a Western shift. Genotek is quite shitty, the same SNP count as V5.

[Jana]

But in his case you added/distracted your values too? Maybe post them I can try with some people.

YES. Try this

North Atlantic + 1.99
Baltic - 3.07
West med -0.05
West Asian + 1.09
East med - 0.19
Red Sea + 0.01
East Asian -0.52
Siberian + 0.67

That's how my results differed. Now other components I score 0 (south Asian, Oceanian etc), so I can't know how these change.
But I often substracted east Asian values from people's amerindian because usually most of them score 0 East Asian so I have nothing to substract and amerindian is still partly EE component.

Ofcourse, I took care their total results end up near 100. All of them get nearly 100 or .40 over it, so it's very close.

[Jana]

looks potentially very useful, but it would be good if you collected more new/old chip pairs and checked if there's really a bias towards Baltic.
Hungarian individuals really can be high Baltic, their oracles will look nothing like an average Hhungarian's, but somebody always has to be on the edge of the cluster.
User Michal is also an east shifted Pole, does he have any ancestry from Kresy? Maybe Baltic ancestry from northern Poland could cause such a shift as well.

Yeah, I agree, ande keep in mind not everyone who has V4 and new tests with less snps will get same differences like I do. But I still think this is pretty useful, it should show something close to reality, at least I think so.
About Roy, I think his ancestry is from further west than Michal if I'm not mistaken. Michal is SE Pole after all, so to me is not strange some of them are as eastern as him.
Roy is at least partly Silesian if I'm not mistaken.

[Leto]

rein's grandmother's results (tested with V5, converted to V3 back in 2018). Too bad he left the forum.

V5:


rein_grandma_V5,29.24,46.46,5.97,3.99,4.63,0.79,0, 2.29,5.5,0.41,0.2,0,0.53

V3:


rein_grandma_V3,27.24,45.24,6.52,5.7,4.92,1.4,0,2. 44,5.51,0.55,0.45,0,0

[Jana]

I did the same for Sharpshooter, who is Irish, and he got North Germans closest with V5 data. Looks good as you can see.


Sharpshooter,49.62,22.93,12.73,9.35,2.95,1.56,0.47 ,0,0.67,0,0,0,0

https://i.ibb.co/0cGWzyV/ss1.png (https://imgbb.com/)

https://i.ibb.co/H79Jwcf/ss2.png (https://imgbb.com/)

[Jana]

rein's grandmother's results (tested with V5, converted to V3 back in 2018). Too bad he left the forum.

V5:


rein_grandma_V5,29.24,46.46,5.97,3.99,4.63,0.79,0, 2.29,5.5,0.41,0.2,0,0.53

V3:


rein_grandma_V3,27.24,45.24,6.52,5.7,4.92,1.4,0,2. 44,5.51,0.55,0.45,0,0


Which chip has more snps V3 or V4?

[vbnetkhio]

Which chip has more snps V3 or V4?

v3, I think it's as good as old FTDNA.

[Leto]

Which chip has more snps V3 or V4?
I think V3 but the difference isn't huge. It goes like this:
1. V3/V4, AncestryDNA, MyHeritage 1.0, FTDNA 1.0
2. MyHeritage 2.0, FTDNA 2.0
3. V5, other less known companies like LivingDNA, Genotek (a Russian company)

[vbnetkhio]

I think V3 but the difference isn't huge. It goes like this:
1. V3/V4, AncestryDNA, MyHeritage 1.0, FTDNA 1.0
2. MyHeritage 2.0, FTDNA 2.0
3. V5, other less known companies like LivingDNA, Genotek (a Russian company)

livingDNA is actually pretty good, around 130k out of 180

[Lucas]

I think V3 but the difference isn't huge. It goes like this:
1. V3/V4, AncestryDNA, MyHeritage 1.0, FTDNA 1.0
2. MyHeritage 2.0, FTDNA 2.0
3. V5, other less known companies like LivingDNA, Genotek (a Russian company)

It is very rare nowadays but I tested some v2 and is of course the best.

[Leto]

livingDNA is actually pretty good, around 130k out of 180
They may have changed the chip 'cause a few years ago they weren't as good.

[knez01]

YES. Try this

North Atlantic + 1.99
Baltic - 3.07
West med -0.05
West Asian + 1.09
East med - 0.19
Red Sea + 0.01
East Asian -0.52
Siberian + 0.67

That's how my results differed. Now other components I score 0 (south Asian, Oceanian etc), so I can't know how these change.
But I often substracted east Asian values from people's amerindian because usually most of them score 0 East Asian so I have nothing to substract and amerindian is still partly EE component.

Ofcourse, I took care their total results end up near 100. All of them get nearly 100 or .40 over it, so it's very close.

In my case, in the converted v3 I get higher Red Sea, South Asian, North Atlantic and West Asian, lower Baltic,West Mediterranean and East Mediterranean. I have just sent my WGS to the lab so once I get the results I will post the differences between the converted and original, although I am completely certain that the converted version is the more correct one. More consistent results in various calculators and more realistic components.

[Ion Basescul]

I can show how mine differ

LivingDNA 58k SNPs so basically like 23andme v5
<google-sheets-html-origin style="color: rgb(0, 0, 0); font-size: medium;"><google-sheets-html-origin>
<tbody>
N_Atlantic
Baltic
West_Med
West_Asian
East_Med
Red_Sea
South_Asian
East_Asian
Siberian
Amerindian
Oceanian
NE_African
Sub-Saharan
Atlant+Balt
Med+WestAsia
EastAsia
India
SSA


25.27
29.82
15.21
10.28
12.81
1.20
0.46
1.82
2.86
0
0.27
0
0
55.09
39.50
4.68
0.46
0

</tbody>
</google-sheets-html-origin></google-sheets-html-origin>
MyHeritage 78k SNPs
<google-sheets-html-origin style="color: rgb(0, 0, 0); font-size: medium;">
<tbody>
N_Atlantic
Baltic
West_Med
West_Asian
East_Med
Red_Sea
South_Asian
East_Asian
Siberian
Amerindian
Oceanian
NE_African
Sub-Saharan
Atlant+Balt
Med+WestAsia
EastAsia
India
SSA


23.53
30.47
15.75
11.25
12.26
1.47
0.66
1.12
2.34
0.10
1.05
0
0
54.00
40.73
3.56
0.66
0

</tbody>
</google-sheets-html-origin>
23andme v4 162k SNPs
<google-sheets-html-origin style="color: rgb(0, 0, 0); font-size: medium;">
<tbody>
N_Atlantic
Baltic
West_Med
West_Asian
East_Med
Red_Sea
South_Asian
East_Asian
Siberian
Amerindian
Oceanian
NE_African
Sub-Saharan
Atlant+Balt
Med+WestAsia
EastAsia
India
SSA


25.50
29.18
15.13
9.99
12.92
1.55
0.65
2.06
2.66
0
0.34
0.02
0
54.68
39.61
4.72
0.65
0

</tbody>
</google-sheets-html-origin>
Merged kit (23andme+MyHeritage+LivingDNA) 171k SNPs
<google-sheets-html-origin style="color: rgb(0, 0, 0); font-size: medium;">
<tbody>
N_Atlantic
Baltic
West_Med
West_Asian
East_Med
Red_Sea
South_Asian
East_Asian
Siberian
Amerindian
Oceanian
NE_African
Sub-Saharan
Atlant+Balt
Med+WestAsia
EastAsia
India
SSA


25.47
29.48
15.06
9.86
12.89
1.53
0.63
1.83
2.85
0
0.40
0
0
54.95
39.34
4.68
0.63
0

</tbody>
</google-sheets-html-origin>

[Jana]

I can show how mine differ

LivingDNA 58k SNPs so basically like 23andme v5
<google-sheets-html-origin style="color: rgb(0, 0, 0); font-size: medium;"><google-sheets-html-origin>
<tbody>
N_Atlantic
Baltic
West_Med
West_Asian
East_Med
Red_Sea
South_Asian
East_Asian
Siberian
Amerindian
Oceanian
NE_African
Sub-Saharan
Atlant+Balt
Med+WestAsia
EastAsia
India
SSA


25.27
29.82
15.21
10.28
12.81
1.20
0.46
1.82
2.86
0
0.27
0
0
55.09
39.50
4.68
0.46
0

</tbody>
</google-sheets-html-origin></google-sheets-html-origin>
MyHeritage 78k SNPs
<google-sheets-html-origin style="color: rgb(0, 0, 0); font-size: medium;">
<tbody>
N_Atlantic
Baltic
West_Med
West_Asian
East_Med
Red_Sea
South_Asian
East_Asian
Siberian
Amerindian
Oceanian
NE_African
Sub-Saharan
Atlant+Balt
Med+WestAsia
EastAsia
India
SSA


23.53
30.47
15.75
11.25
12.26
1.47
0.66
1.12
2.34
0.10
1.05
0
0
54.00
40.73
3.56
0.66
0

</tbody>
</google-sheets-html-origin>
23andme v4 162k SNPs
<google-sheets-html-origin style="color: rgb(0, 0, 0); font-size: medium;">
<tbody>
N_Atlantic
Baltic
West_Med
West_Asian
East_Med
Red_Sea
South_Asian
East_Asian
Siberian
Amerindian
Oceanian
NE_African
Sub-Saharan
Atlant+Balt
Med+WestAsia
EastAsia
India
SSA


25.50
29.18
15.13
9.99
12.92
1.55
0.65
2.06
2.66
0
0.34
0.02
0
54.68
39.61
4.72
0.65
0

</tbody>
</google-sheets-html-origin>
Merged kit (23andme+MyHeritage+LivingDNA) 171k SNPs
<google-sheets-html-origin style="color: rgb(0, 0, 0); font-size: medium;">
<tbody>
N_Atlantic
Baltic
West_Med
West_Asian
East_Med
Red_Sea
South_Asian
East_Asian
Siberian
Amerindian
Oceanian
NE_African
Sub-Saharan
Atlant+Balt
Med+WestAsia
EastAsia
India
SSA


25.47
29.48
15.06
9.86
12.89
1.53
0.63
1.83
2.85
0
0.40
0
0
54.95
39.34
4.68
0.63
0

</tbody>
</google-sheets-html-origin>

Your values seem lot more stable than mine

[J. Ketch]

I don't know if it fits in with your theory, but I have noticed that low SNP British Isles kits (especially 58k) have more North Atlantic and less Baltic, thus more Western shifted.

My original Ancestry data (168k), that I use on here.


1 North_Atlantic 53.96
2 Baltic 24.43
3 West_Med 11.35
4 West_Asian 4.87
5 East_Med 2.57
6 South_Asian 1.3
7 Sub-Saharan 0.75
8 Amerindian 0.45
9 Siberian 0.31

CreodaAncestry,53.96,24.43,11.35,4.87,2.57,0,1.3,0 ,0.31,0.45,0,0,0.75

Target: CreodaAncestry
Distance: 2.0111% / 2.01109448
82.9 West_Scottish
17.1 Norway_South_Central


FTDNA original data (78k)


# Population Percent
1 North_Atlantic 55.63
2 Baltic 24.59
3 West_Med 11.36
4 West_Asian 3.34
5 South_Asian 1.96
6 East_Med 1.91
7 Sub-Saharan 0.6
8 Amerindian 0.56
9 Oceanian 0.04

CreodaFTDNA,55.63,24.59,11.36,3.34,1.91,0,1.96,0,0 ,0.56,0,0,0.6

Target: CreodaFTDNA
Distance: 3.8485% / 3.84847174
87.8 West_Scottish
12.2 Norway_South_Central

My FTDNA data converted to V3 format (182k snps)



1 North_Atlantic 54.62
2 Baltic 24.24
3 West_Med 11.52
4 West_Asian 4.49
5 East_Med 2.4
6 South_Asian 1.39
7 Sub-Saharan 0.67
8 Amerindian 0.61
9 Siberian 0.05

CreodaFTDNAv3,54.62,24.24,11.52,4.49,2.4,0,1.39,0, 0.05,0.61,0,0,0.67

Target: CreodaFTDNAv3
Distance: 2.4614% / 2.46140276
89.3 West_Scottish
10.7 Norway_South_Central


My LivingDNA data converted to V3 (182k snps)


1 North_Atlantic 54.43
2 Baltic 24.24
3 West_Med 11.63
4 West_Asian 4.74
5 East_Med 1.95
6 South_Asian 1.39
7 Amerindian 0.68
8 Siberian 0.54
9 Sub-Saharan 0.39

CreodaLDNA,54.43,24.24,11.63,4.74,1.95,0,1.39,0,0. 54,0.68,0,0,0.39

Target: CreodaLDNA
Distance: 2.0249% / 2.02492074
89.5 West_Scottish
10.5 Norway_South_Central


By contrast my 58k Living DNA original data had 55% North Atlantic and 23.94% Baltic.

Maybe I should use my converted FTDNA or LDNA data instead, they have more SNPs and less SSA, not that I'm complexed or anything :rolleyes:

[Leto]

Maybe I should use my converted FTDNA or LDNA data instead, they have more SNPs and less SSA, not that I'm complexed or anything :rolleyes:
Looks like you bought a lot of tests :D

[J. Ketch]

Looks like you bought a lot of tests :D
3, last one I bought in late 2018 or early 2019. Wouldn't have bought the last FTDNA one if I'd known they had changed the chip unannounced between 2018 and 2019 to lower quality. Serves me right.

[Jana]

^^^^Why do they do that, why did FTDNA, 23andme and MyHeritage lowered quality in their chips? :(

[vbnetkhio]

^^^^Why do they do that, why did FTDNA, 23andme and MyHeritage lowered quality in their chips? :(

it's not lower quality, it just has less overlap with gedmatch calcs, which they don't know or care about.
it's just that all 3 get the chips from the same company (Illumina I think) and a new model came out. FTDNA and MyHeritage ordered a custom version with 20k extra SNPs for better matching with their previous chip version, but it's the same one currently used by 23andme too.

[J. Ketch]

^^^^Why do they do that, why did FTDNA, 23andme and MyHeritage lowered quality in their chips? :(
I guess it's cheaper and quicker for them. DNA sales have been on the decline for the last few years, naturally.

[Ion Basescul]

Maybe I should use my converted FTDNA or LDNA data instead, they have more SNPs and less SSA, not that I'm complexed or anything :rolleyes:

There's no point in converting to guess with imputation for you, considering the data that you already have. Just merge Ancestry with FTDNA and LivingDNA and it should give you the closest result to reality.

[J. Ketch]

There's no point in converting to guess with imputation for you, considering the data that you already have. Just merge Ancestry with FTDNA and LivingDNA and it should give you the closest result to reality.
Is it better to merge the original data or converted?

[Ajeje Brazorf]

My results vary in this way, oddly enough MyHeritage seems to be more accurate:


<colgroup><col><col><col><col><col><col><col><col><col><col><col><col><col><col></colgroup><tbody>
Eurogenes K13
North_Atlantic
Baltic
West_Med
West_Asian
East_Med
Red_Sea
South_Asian
East_Asian
Siberian
Amerindian
Oceanian
Northeast_African
Sub-Saharan


MyHeritage (78,082 SNPs)
19,47
4,01
24,33
14,2
30,06
5,79
0
0,84
0
0
0
1,29
0


Imputed 23andMe V3 (182,231 SNPs)
21,09
3,73
23,9
13,16
30,29
5,38
0
1,47
0
0
0
0,9
0


Difference
-1,62
0,28
0,43
1,04
-0,23
0,41
0
-0,63
0
0
0
0,39
0

</tbody>

[Ion Basescul]

Is it better to merge the original data or converted?

Original of course. Conversion/imputation works by predicting the missing SNPs from those that are already there. It's not guaranteed to be correct.

[knez01]

Is it better to merge the original data or converted?

With the original. The reasoning behind that is the fact that the converted data already has 99%+ genotyping rate on all gedmatch and DIY calculators. The imputation process is extremely reliable but the point of merging is combining tested data to get a better genotyping rate on the calculators. For example on my v5 I get 30% genotyping rate on k13, 20% on MDLP and 25% on Dodecad, which is absolutely horrible. Merging with the converted version wouldn't change your results at all I am pretty sure, because of like I said, maximal genotyping rate.