Similarity between two people using K36

[Peterski]

Use this tool and K36 percentages to compare similarity between two people, for example your parents:

https://gen3553.pagesperso-orange.fr/ADN/comparaison.htm

Between my parents - 86%

https://i.imgur.com/2FpkiUX.png

If you tested your parents, post % similarity between them.

[Brás Garcia de Mascarenhas]

Me and Spaniard 2.0's mother - 84%

Me and Spaniard 2.0's father - 67%

[Jana]

My parents are 77% similar.

[mitalit]

Between my parents - 74%

[Peterski]

I wonder if anyone has more similar parents than 86%.

[Jana]

Stears and me = 73% similarity

[Brás Garcia de Mascarenhas]

Me and my father - 84%

Me and my wife - 78%

[PT Tagus]

Me and Spaniard 2.0's mother - 73%

Me and Spaniard 2.0's father - 63%

[gixajo]

My parents 77%.

https://i.imgur.com/Yc6WWzn.png

[gixajo]

We can calculate it ourselves simply just taking the shared percentage between each components and calculating the total sum.

We can compare different ethnicities/nationalities with this tool too.:)

[gixajo]

And with a simple excel table, we can do exactly the same tool for K13 or any other Gedmatch admixture calculator.

[Brás Garcia de Mascarenhas]

Me and Gaditanian (Andalusian) - 66%

Me and PT Tagus (Portuguese) - 78%

Me and Pedro Ruben (Portuguese) - 79%

Me and Helen (Murcian) - 71%

Me and Ruderico (Portuguese) - 75%

Me and Damião de Góis (Portuguese) - 75%

And the winner, Me and Luso - 87%

I would have done with other members like gixajo or defcon2 but I couldn't find their results on the K36 thread.

[gixajo]

Me and Gaditanian (Andalusian) - 66%

Me and PT Tagus (Portuguese) - 78%

Me and Pedro Ruben (Portuguese) - 79%

Me and Helen (Murcian) - 71%

Me and Ruderico (Portuguese) - 75%

Me and Damião de Góis (Portuguese) - 75%

And the winner, Me and Luso - 87%

I would have done with other members like gixajo or defcon2 but I couldn't find their results on the K36 thread.

I´ve already posted my parents results in this thread.

Mines are:

Amerindian -
Arabian 3.22 Pct
Armenian -
Basque 11.07 Pct
Central_African -
Central_Euro -
East_African -
East_Asian -
East_Balkan 0.79 Pct
East_Central_Asian -
East_Central_Euro 9.6 Pct
East_Med -
Eastern_Euro -
Fennoscandian 0.06 Pct
French 2.25 Pct
Iberian 34.22 Pct
Indo-Chinese -
Italian 10.21 Pct
Malayan -
Near_Eastern -
North_African 2.97 Pct
North_Atlantic 9.34 Pct
North_Caucasian 2.25 Pct
North_Sea 5.53 Pct
Northeast_African -
Oceanian -
Omotic -
Pygmy -
Siberian -
South_Asian -
South_Central_Asian -
South_Chinese -
Volga-Ural -
West_African -
West_Caucasian -
West_Med 8.43 Pct

Where we can observe a few apparent inconsistencies quite evident.

[Brás Garcia de Mascarenhas]

Me and Gixajo - 62%

You are indeed very Basque-shifted, unlike me.

[gixajo]

I cannot get the k36 datasheet in Gedmatch, but I suppose that the similarity map of the k36 that we usually use from this same website, is based on this same comparison of percentages with each reference.

Sorry to reveal the author's trick, but I guess that's the way it is.:rolleyes:

[calxpal]

Me versus my sister: 81%
My parents: 52%

[Peterski]

Me and Roy - 81%

https://i.imgur.com/DTs3EQ7.png

[gixajo]

Me and Gixajo - 62%

You are indeed very Basque-shifted, unlike me.

My Andalusian father score more Basque in K36 than me, and only a 30% less than my "Basque" mother, so well, I don´t know we can give much credit to this.

I was thinking in comparing my Myheritage and my 23AndMe K36, maybe I am under the 90% of similarity with myself.

We can bet the % that I will score with myself...;)

[Peterski]

I was thinking in comparing my Myheritage and my 23AndMe K36, maybe I am under the 90% of similarity with myself.

Yes, it is possible with different raw data files. :)

My scores I posted above are based on a merged raw data file from three companies (FTDNA, 23andMe v5 and Living DNA files merged).

I still need to merge raw data files from two other companies (AncestryDNA and 24Genetics) to get one super-file (5 files merged into 1).

[Peterski]

Even 23andMe v3 vs. 23andMe v5 could be less than 90% similar for the same person, I think so.

You can do Imputation after uploading your raw data to DNA Land and downloading Imputed VCF.


​I was thinking in comparing my Myheritage and my 23AndMe

We can bet the % that I will score with myself...;)

It depends which 23andMe chip it is, and which version of MyHeritage raw data (new or old).

[Axios]

Me and Gixajo: 37%

My k36:

Population
Amerindian -
Arabian 3.74 Pct
Armenian -
Basque 1.1 Pct
Central_African -
Central_Euro -
East_African 1.65 Pct
East_Asian -
East_Balkan 1.42 Pct
East_Central_Asian -
East_Central_Euro -
East_Med 7.37 Pct
Eastern_Euro -
Fennoscandian -
French 1.29 Pct
Iberian 17.32 Pct
Indo-Chinese -
Italian 3.4 Pct
Malayan -
Near_Eastern 11.56 Pct
North_African 28.6 Pct
North_Atlantic 0.26 Pct
North_Caucasian -
North_Sea 3.45 Pct
Northeast_African 14.48 Pct
Oceanian 0.36 Pct
Omotic -
Pygmy -
Siberian -
South_Asian -
South_Central_Asian -
South_Chinese -
Volga-Ural -
West_African 0.38 Pct
West_Caucasian -
West_Med 3.55 Pct


https://i.ibb.co/q57rzZc/gxajo.jpg

[gixajo]

Yes, it is possible with different raw data files. :)

My scores I posted above are based on a merged raw data file from three companies (FTDNA, 23andMe v5 and Living DNA files merged).

I still need to merge raw data files from two other companies (AncestryDNA and 24Genetics) to get one super-file (5 files merged into 1).

What do you call "merged" a mean between results of K36 or a "more complicated" thing as adding all possible SNPs in the same csv/any-format file?

Edit:And yes, I knew it.

[Peterski]

My parents: 52%

New Worlder / Euromutt detected. :cool:


What do you call "merged" a mean between results of K36 or a "more complicated" thing as adding all possible SNPs in the same csv/any-format file?

This.

I merged the raw data files using plink. And then I checked K36 for this merged file.

I also have Global25 coordinates generated from this merged file (based on 3 files).

[calxpal]

Euromutt detected. :cool:

You're right! :eek::cool::033102to:

[gixajo]

Well,here 23andME vs. MyHeritage:

https://i.imgur.com/T8gz1fp.png

[gixajo]

Me and Gixajo: 37%

ct

I am more Arab than you or Nassbean. :rolleyes:

[Peterski]

Well,here 23andME vs. MyHeritage

93% not bad!

[gixajo]

New Worlder / Euromutt detected. :cool:

This.

I merged the raw data files using plink. And then I checked K36 for this merged file.

I also have Global25 coordinates generated from this merged file (based on 3 files).

I would be more interested in doing such a thing with my fathers, because I am actually a mix between North and South, and they are more interesting than me in that aspect.

But they have just been tested with MyHeritage, and my father refuses to take another test.

Is there a tutorial on how to do it with Plink?

I can think of a way to do it, but it would be very slow.

[gixajo]

93% not bad!

It could have been worse.;)

[Peterski]

Is there a tutorial on how to do it with Plink?

I will send you in a PM tomorrow.

[HelloGuys]

93% not bald!

Fixed ;) :cool:

[paradox]

My father and I - 79%

[Komintasavalta]

We can calculate it ourselves simply just taking the shared percentage between each components and calculating the total sum.

Ok, apparently it just calculates the sum of the minimum value on each row. Or at least that gave me 76.87% similarity for your parents, when the calculator gave 77% similarity after rounding:

$ echo '0 0
0 4.94
0 0
14.07 11.66
0 0
0 .92
0 0
0 0
2.08 .94
0 0
6.72 4.94
0 .05
.08 1.24
.42 0
5.26 .29
30.38 34.92
0 0
8.4 16.16
0 0
2.16 0
1.91 4.75
13.48 7.72
0 0
6.64 3.63
0 0
0 .13
0 0
0 .27
0 0
0 0
0 0
0 0
.43 0
0 0
0 .51
7.92 6.92'>gix
$ awk '{s+=$1<$2?$1:$2}END{print s}' gix
76.87

Manhattan distance gives a different result:

$ awk '{x=$1-$2;if(x<0)x=-x;s+=x}END{print s}' gix
46.2

[gixajo]

Ok, apparently it just calculates the sum of the minimum value on each row. Or at least that gave me 76.87% similarity for your parents, when the calculator gave 77% similarity after rounding:

Manhattan distance gives a different result:

$ awk '{x=$1-$2;if(x<0)x=-x;s+=x}END{print s}' gix
46.2

Because what the tool shows is a simple similarity rate result, not an "informed search algorithm" used in computing (I translate literally from Spanish, because I don't know the proper term in English).

[Coastal Elite]

My mother and I - 78%

[Celestia]

My mom and I = 81%

[someonenotyou]

61% for my parents.

[InfamousAngel99]

My parents are 82% similar.
My dad and I are 90% similar.
My mom and I are 81% similar.

I don’t get how my parents are more similar than my mom and I are, lmao, but ok.

[Axios]

My parents are 82% similar.
My dad and I are 90% similar.
My mom and I are 81% similar.

I don’t get how my parents are more similar than my mom and I are, lmao, but ok.Ur adopted

Enviado desde mi Mi 9T Pro mediante Tapatalk

[Peterski]

My parents are 82% similar.

They must have similar ethnic heritage. Your profile says "Euro-mutt", but your parents are as similar to each other as people from the same ethnic group usually are (in this tool).

[InfamousAngel99]

They must have similar ethnic heritage. Your profile says "Euro-mutt", but your parents are as similar to each other as people from the same ethnic group usually are (in this tool).

It’s because they share some ethnic origins, and then have some that the other doesn’t. Within the past 4 generations:

Shared: German, English, Irish, Scottish
Mom only: Croatian, Norwegian, Polish
Dad only: French

[Peterski]

BUMP