Confused about GEDmatch calculator results [Archive] - The Apricity Forum: A European Cultural Community

View Full Version : Confused about GEDmatch calculator results

brjhh2001

10-03-2021, 12:49 PM

Hi guys! :D

According to Ancestry, my DNA is
47% England and Northwestern Europe
27% Ireland
26% Scotland
These roughly fall in line with my documented ancestry.

These are a few of my Gedmatch admixure results. I am a bit confused as to how to interpret them? They seem quite different to my known ancestry. Any help would be greatly appreciated. Cheers! :)

Eurogenes EUtest V2 K15 Oracle results:

Admix Results (sorted):

# Population Percent
1 North_Sea 41.84
2 Atlantic 25.24
3 West_Med 10.79
4 Baltic 9.07
5 Eastern_Euro 5.31
6 East_Med 4.73
7 Southeast_Asian 1.36
8 Sub-Saharan 0.78
9 West_Asian 0.46
10 Siberian 0.23
11 Amerindian 0.11
12 Oceanian 0.09

Single Population Sharing:

# Population (source) Distance
1 Orcadian 7.2
2 Southwest_English 7.99
3 West_Norwegian 8.03
4 North_Dutch 8.12
5 Southeast_English 8.25
6 West_Scottish 8.73
7 Norwegian 8.93
8 West_German 9
9 Irish 9.08
10 Danish 9.18
11 Swedish 10.28
12 North_German 11.45
13 South_Dutch 11.6
14 French 13.41
15 North_Swedish 14.04
16 East_German 17.54
17 Spanish_Galicia 19.86
18 Southwest_Finnish 20.31
19 Spanish_Cataluna 20.48
20 Hungarian 20.94

Mixed Mode Population Sharing:

# Primary Population (source) Secondary Population (source) Distance
1 80% West_Norwegian + 20% Spanish_Murcia @ 5.86
2 78.8% West_Norwegian + 21.2% Portuguese @ 5.88
3 77.6% West_Norwegian + 22.4% Spanish_Galicia @ 5.88
4 89.7% West_Norwegian + 10.3% Sardinian @ 5.9
5 78.5% West_Norwegian + 21.5% Spanish_Cataluna @ 5.98
6 80.5% West_Norwegian + 19.5% Spanish_Extremadura @ 6
7 79.6% West_Norwegian + 20.4% Spanish_Castilla_Y_Leon @ 6
8 82.5% West_Norwegian + 17.5% Spanish_Aragon @ 6.1
9 81.4% West_Norwegian + 18.6% Spanish_Cantabria @ 6.17
10 82.4% West_Norwegian + 17.6% Spanish_Valencia @ 6.18
11 82.6% West_Norwegian + 17.4% Spanish_Castilla_La_Mancha @ 6.2
12 83.6% West_Norwegian + 16.4% Spanish_Andalucia @ 6.22
13 64.6% Orcadian + 35.4% West_German @ 6.27
14 70.2% West_Norwegian + 29.8% French @ 6.27
15 83.7% West_Norwegian + 16.3% North_Italian @ 6.37
16 50.4% Southwest_English + 49.6% West_Norwegian @ 6.49
17 94.1% Orcadian + 5.9% Sardinian @ 6.55
18 83.6% West_Norwegian + 16.4% Southwest_French @ 6.56
19 57.2% West_Norwegian + 42.8% West_German @ 6.61
20 94.3% Orcadian + 5.7% Libyan_Jewish @ 6.62

Eurogenes K13 Oracle results:

Admix Results (sorted):

# Population Percent
1 North_Atlantic 52.09
2 Baltic 21.17
3 West_Med 13.62
4 East_Med 7.67
5 West_Asian 2
6 East_Asian 1.49
7 Sub-Saharan 0.62
8 Amerindian 0.42
9 Northeast_African 0.34
10 Siberian 0.33
11 Oceanian 0.26

Single Population Sharing:

# Population (source) Distance
1 Southeast_English 4.32
2 Southwest_English 5.55
3 Orcadian 6.26
4 West_Scottish 6.76
5 Irish 7.38
6 North_Dutch 7.67
7 Danish 7.85
8 South_Dutch 8
9 West_German 9.17
10 North_German 9.49
11 Norwegian 9.65
12 French 11.33
13 Swedish 11.9
14 Austrian 15.88
15 East_German 16.68
16 North_Swedish 17.22
17 Spanish_Cataluna 17.23
18 Spanish_Castilla_Y_Leon 18.35
19 Southwest_French 18.89
20 Spanish_Cantabria 19.11

Mixed Mode Population Sharing:

# Primary Population (source) Secondary Population (source) Distance
1 92.6% Southeast_English + 7.4% Spanish_Murcia @ 4.03
2 93% Southeast_English + 7% Spanish_Aragon @ 4.05
3 93.2% Southeast_English + 6.8% Spanish_Valencia @ 4.08
4 92.8% Southeast_English + 7.2% Spanish_Castilla_Y_Leon @ 4.09
5 92.7% Southeast_English + 7.3% Spanish_Cataluna @ 4.12
6 94.1% Southeast_English + 5.9% Spanish_Castilla_La_Mancha @ 4.14
7 94% Southeast_English + 6% Portuguese @ 4.15
8 95.3% Southeast_English + 4.7% French_Basque @ 4.15
9 94.9% Southeast_English + 5.1% Spanish_Andalucia @ 4.16
10 98.9% Southeast_English + 1.1% She @ 4.17
11 98.9% Southeast_English + 1.1% Tujia @ 4.18
12 94.9% Southeast_English + 5.1% Spanish_Extremadura @ 4.18
13 98.9% Southeast_English + 1.1% Miaozu @ 4.18
14 99% Southeast_English + 1% Dai @ 4.18
15 98.9% Southeast_English + 1.1% Lahu @ 4.19
16 99% Southeast_English + 1% Vietnamese @ 4.19
17 94.7% Southeast_English + 5.3% Spanish_Cantabria @ 4.19
18 98% Southeast_English + 2% Algerian @ 4.2
19 98.9% Southeast_English + 1.1% Yizu @ 4.2
20 98.9% Southeast_English + 1.1% Naxi @ 4.2

MDLP K23b Oracle results:

Admix Results (sorted):

# Population Percent
1 European_Hunters_Gatherers 34.91
2 European_Early_Farmers 31.93
3 Caucasian 19.81
4 South_Central_Asian 5.31
5 Ancestral_Altaic 4.57
6 North_African 2.37
7 Australoid 0.6
8 Archaic_Human 0.3
9 Tungus-Altaic 0.18
10 Amerindian 0.01
11 Archaic_African 0.01

Single Population Sharing:

# Population (source) Distance
1 English_Kent_GBR ( ) 2.11
2 English_Cornwall_GBR ( ) 2.45
3 Welsh ( ) 3.23
4 CEU ( ) 3.54
5 British ( ) 3.56
6 North_European ( ) 3.89
7 English ( ) 4.28
8 Irish ( ) 5.34
9 French ( ) 5.36
10 Belgian ( ) 5.64
11 Frisian ( ) 6.1
12 Norwegian_West ( ) 6.81
13 Scottish_Argyll_Bute_GBR ( ) 6.93
14 Norwegian_East ( ) 7.17
15 Orcadian ( ) 7.52
16 Icelandic ( ) 8.16
17 Dutch ( ) 9.13
18 Swede ( ) 9.39
19 South_German ( ) 10.22
20 Dane ( ) 10.3

Mixed Mode Population Sharing:

# Primary Population (source) Secondary Population (source) Distance
1 70.3% Dutch ( ) + 29.7% Spanish_Cantabria_IBS ( ) @ 1.35
2 70.6% Dutch ( ) + 29.4% Spanish_Aragon_IBS ( ) @ 1.42
3 91.8% English_Kent_GBR ( ) + 8.2% Spanish_Canarias_IBS ( ) @ 1.45
4 59% North_European ( ) + 41% French ( ) @ 1.51
5 87.7% English_Kent_GBR ( ) + 12.3% Portugese ( ) @ 1.52
6 82.4% English ( ) + 17.6% Spanish_Valencia_IBS ( ) @ 1.54
7 78.6% English ( ) + 21.4% Spanish_Cataluna_IBS ( ) @ 1.56
8 57.8% French ( ) + 42.2% Norwegian_East ( ) @ 1.63
9 69.2% Dutch ( ) + 30.8% Spanish_Castilla_la_Mancha_IBS ( ) @ 1.66
10 98.5% English_Kent_GBR ( ) + 1.5% Mozabite ( ) @ 1.66
11 84.4% English ( ) + 15.6% Spanish_Cantabria_IBS ( ) @ 1.69
12 67.5% Dutch ( ) + 32.5% Spanish_Valencia_IBS ( ) @ 1.69
13 74.7% Dutch ( ) + 25.3% French_South ( ) @ 1.71
14 81.6% English ( ) + 18.4% Spanish_Andalucia_IBS ( ) @ 1.71
15 83.7% English ( ) + 16.3% Spanish_Castilla_la_Mancha_IBS ( ) @ 1.72
16 84.7% English ( ) + 15.3% Spanish_Aragon_IBS ( ) @ 1.72
17 91% English ( ) + 9% Sardinian ( ) @ 1.72
18 56.4% English ( ) + 43.6% French ( ) @ 1.74
19 91.9% North_European ( ) + 8.1% Sardinian ( ) @ 1.75
20 98.2% English_Kent_GBR ( ) + 1.8% Berber_WGA ( ) @ 1.75

Dodecad K12b Oracle results:

Admix Results (sorted):

# Population Percent
1 Atlantic_Med 41.33
2 North_European 40.89
3 Gedrosia 9.01
4 Caucasus 5.35
5 Northwest_African 1.31
6 East_Asian 0.94
7 Southwest_Asian 0.83
8 Sub_Saharan 0.24
9 South_Asian 0.09

Single Population Sharing:

# Population (source) Distance
1 Kent (1000Genomes) 3.85
2 English (Dodecad) 4.55
3 CEU30 (1000Genomes) 4.56
4 Mixed_Germanic (Dodecad) 4.67
5 Cornwall (1000Genomes) 5.01
6 Dutch (Dodecad) 5.14
7 French (Dodecad) 5.44
8 British (Dodecad) 5.51
9 British_Isles (Dodecad) 5.64
10 French (HGDP) 6.11
11 Irish (Dodecad) 6.93
12 Argyll (1000Genomes) 7.16
13 Orcadian (HGDP) 7.26
14 Orkney (1000Genomes) 7.61
15 German (Dodecad) 11.25
16 Norwegian (Dodecad) 14.55
17 Swedish (Dodecad) 17.11
18 Cataluna (1000Genomes) 18.18
19 Hungarians (Behar) 18.59
20 Galicia (1000Genomes) 19.25

Mixed Mode Population Sharing:

# Primary Population (source) Secondary Population (source) Distance
1 80.6% Dutch (Dodecad) + 19.4% Castilla_Y_Leon (1000Genomes) @ 1.2
2 80.5% Dutch (Dodecad) + 19.5% Spanish (Dodecad) @ 1.23
3 82.2% Dutch (Dodecad) + 17.8% Murcia (1000Genomes) @ 1.27
4 80% Dutch (Dodecad) + 20% Extremadura (1000Genomes) @ 1.34
5 80% Dutch (Dodecad) + 20% Portuguese (Dodecad) @ 1.34
6 83.3% Dutch (Dodecad) + 16.7% Andalucia (1000Genomes) @ 1.37
7 82% Dutch (Dodecad) + 18% Castilla_La_Mancha (1000Genomes) @ 1.44
8 79.6% Dutch (Dodecad) + 20.4% Galicia (1000Genomes) @ 1.48
9 80.7% Dutch (Dodecad) + 19.3% Spaniards (Behar) @ 1.49
10 89.7% British_Isles (Dodecad) + 10.3% Morocco_Jews (Behar) @ 1.5
11 81% Dutch (Dodecad) + 19% Cantabria (1000Genomes) @ 1.51
12 78.7% Dutch (Dodecad) + 21.3% Cataluna (1000Genomes) @ 1.52
13 82.3% Dutch (Dodecad) + 17.7% Aragon (1000Genomes) @ 1.53
14 81.9% Dutch (Dodecad) + 18.1% Valencia (1000Genomes) @ 1.55
15 83.7% Dutch (Dodecad) + 16.3% Canarias (1000Genomes) @ 1.65
16 81% Dutch (Dodecad) + 19% Baleares (1000Genomes) @ 1.71
17 89.9% British_Isles (Dodecad) + 10.1% Sephardic_Jews (Behar) @ 1.71
18 93.2% Kent (1000Genomes) + 6.8% Morocco_Jews (Behar) @ 1.74
19 88.5% British_Isles (Dodecad) + 11.5% Sicilian (Dodecad) @ 1.78
20 85.7% CEU30 (1000Genomes) + 14.3% Canarias (1000Genomes) @ 1.79

Komintasavalta

10-03-2021, 01:51 PM

Through Mantel's test, I found that you can make distances calculated based on K13 more accurate by multiplying a K13 datasheet with an MDS matrix of the FST matrix of distances between the components in K13: https://www.theapricity.com/forum/showthread.php?352879-Does-Dodecad-K12-or-Gedrosia-K12-calculator-make-more-sense&p=7310315&viewfull=1#post7310315. It also makes PCAs generated from a K13 datasheet more realistic. Here's a PCA of you and your 50 closest neighbors in the K13 updated datasheet after multiplying by MDS of FST:

https://i.ibb.co/0jqn0H5/1.png

So you plot between the English and Flemish samples. The colored clusters are based on cutting a hierarchical clustering tree at the height where it has 16 subtrees, but based on the clustering you become a lone outlier in your own cluster. (However it could be because the K13 datasheet is based on population averages, where random individual-level variation is averaged out, so individual samples might tend to cluster differently than population averages.) In the plot above, each population is connected with a line to its two nearest neighbors, but you're connected to the population labeled Afrikaner, because you're its closest neighbor even though it is not in your two closest neighbors.

library(tidyverse)
library(ggforce)
library(ggrepel)

k13=read.csv("https://pastebin.com/raw/UY1Em6qW",r=1) # K13 updated
# k13=read.csv("https://pastebin.com/raw/efn67tkr",r=1) # K13 original

k13=rbind(k13,setNames(read.csv(head=F,r=1,text="brjhh2001,52.09,21.17,13.62,2,7.67,0,0,1.49,0.33,0 .42,0.26,0.34,0.62"),names(k13)))

fst=as.dist(read.csv(text=",North_Atlantic,Baltic,West_Med,West_Asian,East_Me d,Red_Sea,South_Asian,East_Asian,Siberian,Amerindi an,Oceanian,Northeast_African,Sub-Saharan
North_Atlantic,,,,,,,,,,,,,
Baltic,19,,,,,,,,,,,,
West_Med,28,36,,,,,,,,,,,
West_Asian,26,32,36,,,,,,,,,,
East_Med,26,35,28,21,,,,,,,,,
Red_Sea,52,62,50,48,39,,,,,,,,
South_Asian,64,65,76,57,60,82,,,,,,,
East_Asian,114,114,122,110,111,127,76,,,,,,
Siberian,111,111,123,109,112,130,83,56,,,,,
Amerindian,138,137,154,138,144,161,120,113,105,,,,
Oceanian,179,181,187,177,176,191,146,166,177,217,, ,
Northeast_African,122,127,124,116,108,121,113,145, 151,185,203,,
Sub-Saharan,146,150,150,140,135,141,133,164,170,204,22 0,41,",r=1))/1000

mds=cmdscale(as.matrix(fst),ncol(as.matrix(fst))-2)
t=as.matrix(k13/100)%*%mds
# t=as.data.frame(as.matrix(k13)%*%as.matrixfst)
write.table(round(t,6),"k13mds",quote=F,col.names=F,sep=",")

t=t[names(head(sort(as.matrix(dist(t))["brjhh2001",]),50)),]

p0=prcomp(t)
p=adf(p0$x)
pct=paste0(colnames(p0$x)," (",sprintf("%.1f",p0$sdev/sum(p0$sdev)*100),"%)")

i=1
xpc=sym(paste0("PC",i))
ypc=sym(paste0("PC",i+1))

seg=lapply(1:3,function(j)apply(as.matrix(dist(t)) ,1,function(x)unlist(p[names(sort(x)[j]),c(i,i+1)],use.names=F))%>%t%>%cbind(p[,c(i,i+1)]))%>%do.call(rbind,.)%>%setNames(paste0("V",1:4))

p$k=as.factor(cutree(hclust(dist(t)),16))

ggplot(p,aes(!!xpc,!!ypc))+
geom_segment(data=seg,aes(x=V1,y=V2,xend=V3,yend=V 4),color="gray60",size=.15)+
ggforce::geom_mark_hull(aes(color=k,fill=k),concav ity=1000,radius=unit(.15,"cm"),expand=unit(.15,"cm"),alpha=.2,size=.15)+
geom_point(aes(color=k),size=.5)+
# geom_text(aes(color=k),label=rownames(p),size=2,vj ust=-.7)+
ggrepel::geom_text_repel(aes(color=k,label=rowname s(p)),size=2,max.overlaps=Inf,force=2,segment.size =.2,min.segment.length=.2,box.padding=.05)+
labs(x=pct[i],y=pct[i+1])+
scale_x_continuous(breaks=seq(-1,1,.1),expand=expansion(mult=c(.04,.04)))+
scale_y_continuous(breaks=seq(-1,1,.1),expand=expansion(mult=c(.04,.04)))+
scale_fill_manual(values=rainbow_hcl(nlevels(p$k), 90,60))+
scale_color_manual(values=rainbow_hcl(nlevels(p$k) ,90,60))+
theme(
axis.text=element_text(size=6),
axis.text.y=element_text(angle=90,vjust=1,hjust=.5 ),
axis.ticks=element_blank(),
axis.ticks.length=unit(0,"cm"),
axis.title=element_text(size=8),
legend.position="none",
panel.background=element_rect(fill="white"),
panel.border=element_rect(color="gray85",fill=NA,size=.6),
panel.grid.major=element_line(color="gray85",size=.2),
plot.background=element_rect(fill="white")
)

ggsave("1.png",width=5,height=5)

I tried using a version of the K13 datasheet that was multiplied by an MDS matrix of the FST matrix, and I made models reduced to a different number of maximum populations:

$ curl https://pastebin.com/raw/afaMiFSa|tr -d \\r>mix;chmod +x mix;pip3 install cvxpy
[...]
$ for m in {2..6};do t=brjhh;f=k13mds;./mix <(grep -v $t $f) <(grep $t $f) -m$m -s -f1;done
brjhh2001 (.001): 78.5% Dutch_North + 21.5% French_Southwest
brjhh2001 (.001): 75.7% Norwegian + 21.3% French_Basque + 3.0% Ethiopian_Tigray
brjhh2001 (.001): 43.6% Sweden_Götaland + 32.6% Norwegian + 21.1% French_Basque + 2.8% Ethiopian_Tigray
brjhh2001 (.001): 44.6% Sweden_Götaland + 30.4% Norwegian + 13.5% Spanish_Aragon + 9.2% French_Basque + 2.3% Ethiopian_Tigray
brjhh2001 (.000): 47.6% Sweden_Götaland + 24.6% Norwegian + 19.4% French_Basque + 5.4% Spanish_Aragon + 2.2% Ethiopian_Tigray + 0.8% She

With the regular K13 datasheet without multiplication by MDS of FST, you got these models:

$ curl https://pastebin.com/raw/UY1Em6qW>k13orig
$ for m in {2..6};do t=brjhh;./mix <(sed 1s/^pop// k13orig) <(echo brjhh2001,52.09,21.17,13.62,2,7.67,0,0,1.49,0.33,0 .42,0.26,0.34,0.62) -m$m -s -f1;done
brjhh2001 (5.180): 79.0% West_Scottish + 21.0% Swiss_Italian
brjhh2001 (5.111): 46.0% West_Scottish + 33.3% Dutch_North + 20.7% Spanish_Valencia
brjhh2001 (5.109): 45.9% West_Scottish + 33.4% Dutch_North + 16.3% Spanish_Valencia + 4.5% Spanish_Murcia
brjhh2001 (5.109): 45.9% West_Scottish + 33.4% Dutch_North + 16.3% Spanish_Valencia + 4.5% Spanish_Murcia
brjhh2001 (5.109): 45.9% West_Scottish + 33.4% Dutch_North + 16.3% Spanish_Valencia + 4.5% Spanish_Murcia

(This uses my modified version of michal3141's convex optimization script (https://github.com/michal3141/g25), which produces virtually identical results to Vahaduo.)

J. Ketch

10-03-2021, 02:09 PM

You have some non British Isles ancestry in your genealogy?

brjhh2001

10-03-2021, 02:15 PM

Wow, this is great, thank you so much!
I find these results very interesting. I am a bit perplexed as to the seemingly quite high Germanic and Nordic (I assume I'm interpreting these correctly) and the relationship to the Afrikaner population. I do not know of any recent Germanic or Scandinavian ancestry in my tree whatsoever. Any thoughts on why this might be the case? The rest makes sense, the K13 x MDS results pick up my African ancestry through my Cape Verdean heritage which I haven't seen any others pick up thus far.
Thanks again, I am genuinely amazed by your awesome work :D

J. Ketch

10-03-2021, 02:16 PM

brjhh2001

10-03-2021, 02:17 PM

On paper, I have a 4th or 5th great grandmother who was French and I have worked out via DNA matches that I have some Cape Verdean heritage but the rest is British Isles as far as I know :)

J. Ketch

10-03-2021, 02:35 PM

On paper, I have a 4th or 5th great grandmother who was French and I have worked out via DNA matches that I have some Cape Verdean heritage but the rest is British Isles as far as I know :)
How you plot in a regular PCA:

https://i.postimg.cc/2mTQ6sRc/Vahaduo-Custom-PCA-5.png

You are Atlantic/'Celtic' shifted, towards NW France, not between SE England and Flanders. But you are certainly within the range of scores for fully British Isles people, so there's no mystery.

J. Ketch

10-03-2021, 02:49 PM

brjhh2001

10-03-2021, 03:21 PM

Your closest updated populations in K13:

Distance to: brjhh2001
6.11768747 Cornish
6.40810424 English_Midlands
6.43154725 English
6.43321848 English_Southwest
6.45328598 English_Southeast
6.65743194 French_Northwest
6.74178760 English_North
7.07091932 Welsh
7.16372808 Orcadian
7.29102188 Dutch
7.29562197 Scottish_East
7.40179708 Scottish_North_Highlands
7.43983871 Dutch_Central
7.51548402 Scottish_Northeast
7.60157878 Scottish_Southwest
7.71430489 Scottish
7.76970398 West_Scottish
7.79719821 Irish_Leinster
7.92718740 Dutch_South
7.99906870 Flemish

Interesting! This does align with my known ancestry on paper. My great grandfather was 100% Cornish and I would have thought through him and my plentiful Irish and Scottish (mainly Irish) ancestors on both sides, I'd be primarily Celtic. That's exactly why these high Germanic and Nordic percentages elsewhere are so puzzling to me. I could understand some minor Norwegian via Scotland but otherwise I'm lost.

Komintasavalta

10-03-2021, 06:22 PM

If you calculate K13 distances the regular way without accounting for FST, then he's closer to Maris than to Sardinians, and he's closer to Southwest_Finnish than to South_Polish, because not enough weight is given to differences in the level of Mongoloid admixture. And he's also closer to most Spanish populations than to Hungarians. And he's closer to many SSAs than to East Asians and Southeast Asians, because for example the distance between the North_Atlantic and Sub-Saharan components is treated as equal to the distance between the North_Atlantic and East_Asian components.

https://i.ibb.co/jrJgFfr/1.png

How you plot in a regular PCA:

https://i.postimg.cc/2mTQ6sRc/Vahaduo-Custom-PCA-5.png

You are Atlantic/'Celtic' shifted, towards NW France, not between SE England and Flanders. But you are certainly within the range of scores for fully British Isles people, so there's no mystery.

It depends on what populations you include. Here's another PCA made using my method of multiplying by MDS of FST, where now he plots close to French_Northwest:

https://i.ibb.co/DL1vHqB/2.png

But maybe my method isn't that good for finding the closest populations for Northwestern Europeans, because it amplifies the noise-level admixture of non-European components. It made the OP closest to Dutch_South and Flemish:

K13 distance to brjhh2001 (multiplied by MDS of FST):
.124 Dutch_South
.150 Flemish
.152 French_Northwest
.161 German_West
.161 Dutch
.161 English_Southeast
.164 English_Midlands
.169 English
.175 Dutch_Central
.175 Belgian
.184 English_Southwest
.186 Welsh
.187 English_North
.196 German
.202 Scottish_North_Highlands
K13 distance to brjhh2001 (not accounting for FST):
6.118 Cornish
6.408 English_Midlands
6.432 English
6.433 English_Southwest
6.453 English_Southeast
6.657 French_Northwest
6.742 English_North
7.071 Welsh
7.164 Orcadian
7.291 Dutch
7.296 Scottish_East
7.402 Scottish_North_Highlands
7.440 Dutch_Central
7.515 Scottish_Northeast
7.602 Scottish_Southwest

I think my methods puts him relatively close to Dutch_South, because they have similar levels of the last six components, which have the highest FST distance to Europeans:

North_AtlanticBalticWest_MedWest_AsianEast_MedRed_ SeaSouth_AsianEast_AsianSiberianAmerindianOceanian Northeast_AfricanSub-Saharan
Dutch_South45.9923.4115.115.155.601.301.030.360.54 0.520.290.340.30
brjhh200152.0921.1713.6227.67001.490.330.420.260.3 40.62

brjhh2001

10-04-2021, 12:37 AM

If you calculate K13 distances the regular way without accounting for FST, then he's closer to Maris than to Sardinians, and he's closer to Southwest_Finnish than to South_Polish, because not enough weight is given to differences in the level of Mongoloid admixture. And he's also closer to most Spanish populations than to Hungarians. And he's closer to many SSAs than to East Asians and Southeast Asians, because for example the distance between the North_Atlantic and Sub-Saharan components is treated as equal to the distance between the North_Atlantic and East_Asian components.

https://i.ibb.co/jrJgFfr/1.png

It depends on what populations you include. Here's another PCA made using my method of multiplying by MDS of FST, where now he plots close to French_Northwest:

https://i.ibb.co/DL1vHqB/2.png

But maybe my method isn't that good for finding the closest populations for Northwestern Europeans, because it amplifies the noise-level admixture of non-European components. It made the OP closest to Dutch_South and Flemish:

K13 distance to brjhh2001 (multiplied by MDS of FST):
.124 Dutch_South
.150 Flemish
.152 French_Northwest
.161 German_West
.161 Dutch
.161 English_Southeast
.164 English_Midlands
.169 English
.175 Dutch_Central
.175 Belgian
.184 English_Southwest
.186 Welsh
.187 English_North
.196 German
.202 Scottish_North_Highlands
K13 distance to brjhh2001 (not accounting for FST):
6.118 Cornish
6.408 English_Midlands
6.432 English
6.433 English_Southwest
6.453 English_Southeast
6.657 French_Northwest
6.742 English_North
7.071 Welsh
7.164 Orcadian
7.291 Dutch
7.296 Scottish_East
7.402 Scottish_North_Highlands
7.440 Dutch_Central
7.515 Scottish_Northeast
7.602 Scottish_Southwest

I think my methods puts him relatively close to Dutch_South, because they have similar levels of the last six components, which have the highest FST distance to Europeans:

North_AtlanticBalticWest_MedWest_AsianEast_MedRed_ SeaSouth_AsianEast_AsianSiberianAmerindianOceanian Northeast_AfricanSub-Saharan
Dutch_South45.9923.4115.115.155.601.301.030.360.54 0.520.290.340.30
brjhh200152.0921.1713.6227.67001.490.330.420.260.3 40.62

Hmm, I'm not aware of any Dutch ancestry in my family tree so these have me slightly confused. The closeness to Northwest France makes a lot of sense as my Cornish DNA would resemble Bretons' quite closely, I would assume. The Dutch and Flemish is interesting. I have read elsewhere that the South Dutch K13 group is a mixture of Germanic and Celtic which I figure could make sense in terms of resembling my overall DNA.

Grace O'Malley

10-04-2021, 01:27 AM

Interesting! This does align with my known ancestry on paper. My great grandfather was 100% Cornish and I would have thought through him and my plentiful Irish and Scottish (mainly Irish) ancestors on both sides, I'd be primarily Celtic. That's exactly why these high Germanic and Nordic percentages elsewhere are so puzzling to me. I could understand some minor Norwegian via Scotland but otherwise I'm lost.

If you spend more time on genetics you would see that being Celtic is not genetic. There is no Celtic cluster for instance. All Isles populations when using calculators like Eurogenes K13 are close to Scandinavians and Dutch because they are all Northwestern Europeans. Using the updated K13 these are my closest populations.

Distance to: Grace(irish)
2.37583669 Icelandic
3.03324908 Scottish_Gaidhealtachd
3.90000000 Irish_Connacht
3.92126255 Scottish
3.97353747 Dutch_North
3.99221743 Scottish_Southwest
4.02823783 West_Scottish
4.04991358 Norwegian
4.11541006 Irish
4.25025882 Irish_Munster
4.25540832 Scottish_Northeast
4.25605451 Orcadian
4.35045975 Scottish_East
4.41269759 Irish_Leinster
4.50302121 Scottish_North_Highlands
4.52268725 Irish_Ulster
4.67891013 Denmark
4.67923071 Dutch_Central
4.86428823 Norway_South_Central
4.90855376 English_North
5.16241223 Welsh
5.19384251 Dutch
5.25249465 Sweden_Götaland
5.75115641 German_Northwest
5.81993127 English

Target: Grace(irish)
Distance: 1.9009% / 1.90091955
69.6 Icelandic
19.6 West_Scottish
9.3 Irish_Ulster
1.3 Dargin
0.2 Karitiana

So your results look more typical for an English person.

Grace O'Malley

10-04-2021, 02:29 AM

This is what you get with the updated K13.

Distance to: brjhh2001
6.11768747 Cornish
6.40810424 English_Midlands
6.43154725 English
6.43321848 English_Southwest
6.45328598 English_Southeast
6.65743194 French_Northwest
6.74178760 English_North
7.07091932 Welsh
7.16372808 Orcadian
7.29102188 Dutch
7.29562197 Scottish_East
7.40179708 Scottish_North_Highlands
7.43983871 Dutch_Central
7.51548402 Scottish_Northeast
7.60157878 Scottish_Southwest
7.71430489 Scottish
7.76970398 West_Scottish
7.79719821 Irish_Leinster
7.92718740 Dutch_South
7.99906870 Flemish
8.31417464 Belgian
8.32517868 Dutch_North
8.40098209 Irish_Munster
8.43306587 Irish
8.83692820 Irish_Ulster

Target: brjhh2001
Distance: 5.1072% / 5.10719780
47.2 West_Scottish
32.1 Dutch_North
18.7 Spanish_Valencia
1.8 Spanish_Murcia
0.2 Dai

Distance to: brjhh2001
5.17987894 21.00% Swiss_Italian + 79.00% West_Scottish
5.26134498 77.60% Dutch_North + 22.40% Spanish_Aragon
5.34632914 82.00% Orcadian + 18.00% Spanish_Murcia
5.36055902 20.60% Spanish_Valencia + 79.40% West_Scottish
5.37494299 82.00% Orcadian + 18.00% Spanish_Valencia
5.38760283 20.60% Spanish_Murcia + 79.40% West_Scottish
5.47226549 82.80% Orcadian + 17.20% Spanish_Aragon
5.48921241 22.80% Spanish_Cataluna + 77.20% West_Scottish
5.49189751 80.20% Orcadian + 19.80% Spanish_Cataluna
5.49490834 81.20% Orcadian + 18.80% Spanish_Castilla_Y_Leon
5.53992053 21.40% Spanish_Castilla_Y_Leon + 78.60% West_Scottish
5.55118489 81.80% Dutch_Central + 18.20% Spanish_Aragon
5.56193099 77.80% Dutch_North + 22.20% Spanish_Valencia
5.56788979 82.80% Orcadian + 17.20% Swiss_Italian
5.57139255 19.80% Swiss-Italian2 + 80.20% West_Scottish
5.57615464 77.80% Dutch_North + 22.20% Spanish_Murcia
5.59586009 20.40% Portuguese + 79.60% West_Scottish
5.61337904 83.00% Orcadian + 17.00% Spanish
5.63313847 90.00% Cornish + 10.00% Spanish_Murcia
5.64432599 82.60% Orcadian + 17.40% Portuguese
5.65695751 90.20% Cornish + 9.80% Spanish_Valencia
5.66410306 90.60% Cornish + 9.40% Spanish_Aragon
5.67625490 83.40% Orcadian + 16.60% Spanish_Castilla_La_Mancha
5.68286422 16.00% Italian_Lombardy + 84.00% West_Scottish
5.69975052 75.40% Dutch_North + 24.60% Spanish_Cataluna