PCA for PuntDNAL k15

[MysteriousWays]

I am willing to start constructing a PCA for PuntDNAL k15. I know it gives the best results for me out there. Please post scores; it will probably be a day or so before I get back to you as I still have to construct the initial scores. But please post your scores and I will get them on a PCA in the next day or two.

[Hadouken]

thanks

edit : initial order

S_Indian 5.44
Mediterranean 26.30
Siberian 1.41
Wht_Nile_River -
Amerindian 1.38
S_African 0.29
E_Asian 0.14
Caucasian 42.94
NE_European 7.75
Omo_River -
W_African 0.63
Horn_Of_Africa 3.21
Oceanian -
Beringian -
SW_Asian 10.53


sorted :

# Population Percent

1 Caucasian 42.94
2 Mediterranean 26.3
3 SW_Asian 10.53
4 NE_European 7.75
5 S_Indian 5.44
6 Horn_Of_Africa 3.21
7 Siberian 1.41
8 Amerindian 1.38
9 W_African 0.63
10 S_African 0.29
11 E_Asian 0.14

[MysteriousWays]

If you can please put it in the initial order that it comes out in GEDMatch, that will also be much faster for me to enter. Thank you.

[magicalM]

Thank you very much in advance!
#PopulationPercent
1NE_European55.6
2Mediterranean26.53
3Caucasian12.62
4SW_Asian2.69
5S_Indian0.57
6Omo_River0.57
7Beringian0.45
8W_African0.4
9Oceanian0.3
10Horn_Of_Africa0.28

Sent from my LG-H815 using Tapatalk

[RN97]

thanks m8


Population
S_Indian 1.66
Mediterranean 32.00
Siberian 1.29
Wht_Nile_River -
Amerindian 0.11
S_African -
E_Asian -
Caucasian 13.81
NE_European 47.61
Omo_River 0.50
W_African -
Horn_Of_Africa -
Oceanian -
Beringian 0.93
SW_Asian 2.07






It gives me pretty wrong oracles though lol
# Primary Population (source) Secondary Population (source) Distance
1 89% South_German + 11% Turk_Istanbul @ 1.76
2 87.8% Utahn_White + 12.2% Turk_Istanbul @ 1.79
3 89.4% South_German + 10.6% Azerbaijani @ 1.84
4 88.2% Utahn_White + 11.8% Azerbaijani @ 1.85
5 86.7% Utahn_White + 13.3% Turk_Kayseri @ 1.87
6 88.1% South_German + 11.9% Turk_Kayseri @ 1.92
7 89% South_German + 11% Kurdish @ 2
8 88.5% South_German + 11.5% Kumyk @ 2.01
9 90.7% South_German + 9.3% Abkhasian @ 2.01
10 89.1% South_German + 10.9% North_Ossetian @ 2.03
11 91.4% South_German + 8.6% Georgian @ 2.04
12 87.8% Utahn_White + 12.2% Kurdish @ 2.04
13 90% French + 10% Tadjik @ 2.07
14 89.7% South_German + 10.3% Armenian @ 2.1
15 90.4% French + 9.6% Lezgin @ 2.1
16 89.8% South_German + 10.2% Iranian @ 2.12
17 67.8% Serbian + 32.2% French @ 2.14
18 88.6% Utahn_White + 11.4% Armenian @ 2.15
19 88.6% South_German + 11.4% Balkar @ 2.16
20 90.1% South_German + 9.9% Turk_Trabzon @ 2.16

[Rius]

Thanks in advance


# Population Percent
1 Caucasian 33.77
2 Mediterranean 26.42
3 NE_European 19.32
4 SW_Asian 10.1
5 E_Asian 5.96
6 S_African 2.51
7 Siberian 1.09
8 Horn_Of_Africa 0.55
9 Beringian 0.29

[Gangrel]

S_Indian 3.32
Mediterranean 24.86
Siberian 4.02
Wht_Nile_River 0.14
Amerindian 1.44
S_African 0.25
E_Asian 1.61
Caucasian 44.72
NE_European 9.33
Omo_River 0.51
W_African 0.67
Horn_Of_Africa 1.02
Oceanian -
Beringian 0.90
SW_Asian 7.21

Thanks

[MysteriousWays]

thanks m8


Population
S_Indian 1.66
Mediterranean 32.00
Siberian 1.29
Wht_Nile_River -
Amerindian 0.11
S_African -
E_Asian -
Caucasian 13.81
NE_European 47.61
Omo_River 0.50
W_African -
Horn_Of_Africa -
Oceanian -
Beringian 0.93
SW_Asian 2.07






It gives me pretty wrong oracles though lol
# Primary Population (source) Secondary Population (source) Distance
1 89% South_German + 11% Turk_Istanbul @ 1.76
2 87.8% Utahn_White + 12.2% Turk_Istanbul @ 1.79
3 89.4% South_German + 10.6% Azerbaijani @ 1.84
4 88.2% Utahn_White + 11.8% Azerbaijani @ 1.85
5 86.7% Utahn_White + 13.3% Turk_Kayseri @ 1.87
6 88.1% South_German + 11.9% Turk_Kayseri @ 1.92
7 89% South_German + 11% Kurdish @ 2
8 88.5% South_German + 11.5% Kumyk @ 2.01
9 90.7% South_German + 9.3% Abkhasian @ 2.01
10 89.1% South_German + 10.9% North_Ossetian @ 2.03
11 91.4% South_German + 8.6% Georgian @ 2.04
12 87.8% Utahn_White + 12.2% Kurdish @ 2.04
13 90% French + 10% Tadjik @ 2.07
14 89.7% South_German + 10.3% Armenian @ 2.1
15 90.4% French + 9.6% Lezgin @ 2.1
16 89.8% South_German + 10.2% Iranian @ 2.12
17 67.8% Serbian + 32.2% French @ 2.14
18 88.6% Utahn_White + 11.4% Armenian @ 2.15
19 88.6% South_German + 11.4% Balkar @ 2.16
20 90.1% South_German + 9.9% Turk_Trabzon @ 2.16

I think it works best for people of mostly Mediterranean origins to south Asia, but we'll see

[firemonkey]

Thanks.

S_Indian 0.18
Mediterranean 28.95
Siberian 0.54
Wht_Nile_River 1.22
Amerindian 0.63
S_African -
E_Asian -
Caucasian 8.08
NE_European 57.08
Omo_River 0.75
W_African -
Horn_Of_Africa -
Oceanian -
Beringian 0.57
SW_Asian 2.00

[MysteriousWays]

I am starting with Europeans first for now as the European averages are now fully entered in. I will get in West Asian, Middle Eastern, and other samples in this evening.

[MysteriousWays]

Thank you very much in advance!
#PopulationPercent
1NE_European55.6
2Mediterranean26.53
3Caucasian12.62
4SW_Asian2.69
5S_Indian0.57
6Omo_River0.57
7Beringian0.45
8W_African0.4
9Oceanian0.3
10Horn_Of_Africa0.28

Sent from my LG-H815 using Tapatalk

Some samples need major corrections in their averages I think, but here is yours. Seems decent-ish I suppose.

https://s5.postimg.org/nuh9n7hgn/Magical_M.png

[MysteriousWays]

thanks m8


Population
S_Indian 1.66
Mediterranean 32.00
Siberian 1.29
Wht_Nile_River -
Amerindian 0.11
S_African -
E_Asian -
Caucasian 13.81
NE_European 47.61
Omo_River 0.50
W_African -
Horn_Of_Africa -
Oceanian -
Beringian 0.93
SW_Asian 2.07






It gives me pretty wrong oracles though lol
# Primary Population (source) Secondary Population (source) Distance
1 89% South_German + 11% Turk_Istanbul @ 1.76
2 87.8% Utahn_White + 12.2% Turk_Istanbul @ 1.79
3 89.4% South_German + 10.6% Azerbaijani @ 1.84
4 88.2% Utahn_White + 11.8% Azerbaijani @ 1.85
5 86.7% Utahn_White + 13.3% Turk_Kayseri @ 1.87
6 88.1% South_German + 11.9% Turk_Kayseri @ 1.92
7 89% South_German + 11% Kurdish @ 2
8 88.5% South_German + 11.5% Kumyk @ 2.01
9 90.7% South_German + 9.3% Abkhasian @ 2.01
10 89.1% South_German + 10.9% North_Ossetian @ 2.03
11 91.4% South_German + 8.6% Georgian @ 2.04
12 87.8% Utahn_White + 12.2% Kurdish @ 2.04
13 90% French + 10% Tadjik @ 2.07
14 89.7% South_German + 10.3% Armenian @ 2.1
15 90.4% French + 9.6% Lezgin @ 2.1
16 89.8% South_German + 10.2% Iranian @ 2.12
17 67.8% Serbian + 32.2% French @ 2.14
18 88.6% Utahn_White + 11.4% Armenian @ 2.15
19 88.6% South_German + 11.4% Balkar @ 2.16
20 90.1% South_German + 9.9% Turk_Trabzon @ 2.16

It's not calculating you well, but here...

https://s5.postimg.org/we0nkypt3/RN97.png

[Alessio]

So everyone nowadays is making PCA's on TA - cool!

Me

Mediterranean 35.35
Siberian 0.16
E_Asian 0.46
Caucasian 18.88
NE_European 37.52
Horn_Of_Africa 1.87
Oceanian 0.18
Beringian 0.39
SW_Asian 5.18

North Dutch 1

S_Indian 0.30
Mediterranean 27.62
Wht_Nile_River 0.41
Amerindian 0.29
Caucasian 11.25
NE_European 58.46
Oceanian 0.22
Beringian 0.71
SW_Asian 0.74

Naples

Mediterranean 40.65
Siberian 0.37
Caucasian 23.29
NE_European 22.46
Horn_Of_Africa 3.02
Oceanian 0.13
SW_Asian 10.05

[Kelmendasi]

Thanks!

Population
S_Indian -
Mediterranean 39.46
Siberian -
Wht_Nile_River -
Amerindian 0.41
S_African -
E_Asian -
Caucasian 17.80
NE_European 34.22
Omo_River -
W_African -
Horn_Of_Africa 0.94
Oceanian -
Beringian -
SW_Asian 7.17

[MysteriousWays]

I'm going to take a pause from calculating more for a bit. I have to enter in more reference populations to let this do its thing properly. I should have them up later tonight.

[MysteriousWays]

Thanks.

S_Indian 0.18
Mediterranean 28.95
Siberian 0.54
Wht_Nile_River 1.22
Amerindian 0.63
S_African -
E_Asian -
Caucasian 8.08
NE_European 57.08
Omo_River 0.75
W_African -
Horn_Of_Africa -
Oceanian -
Beringian 0.57
SW_Asian 2.00

Please confirm...Wht_Nile_River is 1.22? If this is not accurate, it seems to be throwing off the PCA significantly.

[Sikeliot]

I don't understand why the Portuguese plot where they do on that chart.

[MysteriousWays]

New one for MagicalM (should be better)
https://s5.postimg.org/qr4anhnaf/magical_M2.png

[MysteriousWays]

I don't understand why the Portuguese plot where they do on that chart.

I just am making new graphs, but the PCA for this is very sensitive to tiny changes. So one example...Portugal has a "Horn_Africa" score of 4.8,which is among the highest in Europe (other than Sicily and Jews). Personally, I think the sample averages are crap so if someone can get me revised ones that would be great. Even a tiny change on Horn_Africa will make a major shift in these graphs, so that seems to be what is happening.

[MysteriousWays]

RN97, Here is a revised one.

https://s5.postimg.org/s7ft5mq7b/RN972.png

[MysteriousWays]

Alessio....

https://s5.postimg.org/jrq8o4nc7/Alessio_v2.png

[MysteriousWays]

Thanks!

Population
S_Indian -
Mediterranean 39.46
Siberian -
Wht_Nile_River -
Amerindian 0.41
S_African -
E_Asian -
Caucasian 17.80
NE_European 34.22
Omo_River -
W_African -
Horn_Of_Africa 0.94
Oceanian -
Beringian -
SW_Asian 7.17

Here you go, sir...

https://s5.postimg.org/6dd5s3go7/Kelmendasi.png

[MysteriousWays]

S_Indian 3.32
Mediterranean 24.86
Siberian 4.02
Wht_Nile_River 0.14
Amerindian 1.44
S_African 0.25
E_Asian 1.61
Caucasian 44.72
NE_European 9.33
Omo_River 0.51
W_African 0.67
Horn_Of_Africa 1.02
Oceanian -
Beringian 0.90
SW_Asian 7.21

Thanks

Here you go....I changed the dimensionality to separate out MENA from Europe more

https://s5.postimg.org/4e15efbqf/Eggyolk_2.png

[MysteriousWays]

thanks

edit : initial order

S_Indian 5.44
Mediterranean 26.30
Siberian 1.41
Wht_Nile_River -
Amerindian 1.38
S_African 0.29
E_Asian 0.14
Caucasian 42.94
NE_European 7.75
Omo_River -
W_African 0.63
Horn_Of_Africa 3.21
Oceanian -
Beringian -
SW_Asian 10.53


sorted :

# Population Percent

1 Caucasian 42.94
2 Mediterranean 26.3
3 SW_Asian 10.53
4 NE_European 7.75
5 S_Indian 5.44
6 Horn_Of_Africa 3.21
7 Siberian 1.41
8 Amerindian 1.38
9 W_African 0.63
10 S_African 0.29
11 E_Asian 0.14

Here you are. Plot neatly next to Kurds.https://s5.postimg.org/hse8agief/All-_In.png

Second dimension map
https://s5.postimg.org/jb9mffoyv/All-_In2.png

[MysteriousWays]

Thanks in advance


# Population Percent
1 Caucasian 33.77
2 Mediterranean 26.42
3 NE_European 19.32
4 SW_Asian 10.1
5 E_Asian 5.96
6 S_African 2.51
7 Siberian 1.09
8 Horn_Of_Africa 0.55
9 Beringian 0.29

Here you are...Basically showing up in the Azeri range...

https://s5.postimg.org/pmetvuq7b/Rius.png

[MysteriousWays]

BTW, there is a third dimension that should be included with these, which pulls away MENA and South Asians from everyone else, so I may add a few plots next to yours. This does not plot all that well on two dimensions.

[firemonkey]

Please confirm...Wht_Nile_River is 1.22? If this is not accurate, it seems to be throwing off the PCA significantly.

I copied and pasted the results(23andMe data) so it is correct.

[MysteriousWays]

I copied and pasted the results(23andMe data) so it is correct.

Thank you. I think I have it under control now. I will be modifying your results. It was being a bit weird with the principal components it was deriving (it was not what I would consider to be intuitive, that's for sure).

[MysteriousWays]

I copied and pasted the results(23andMe data) so it is correct.

OK....Here are your results.

https://s5.postimg.org/bwkan233b/firemonkey2.png

[Kelmendasi]

OK....Here are your results.

https://s5.postimg.org/bwkan233b/firemonkey2.png
Could you please add me in this dimension as it seems more accurate? For example Albanians are plot north of Greeks as it should be on this dimension but on the other one we didn't

[Hadouken]

Here you are. Plot neatly next to Kurds.https://s5.postimg.org/hse8agief/All-_In.png

Second dimension map
https://s5.postimg.org/jb9mffoyv/All-_In2.png

thank you

I cant read much in the second map . any chance you can fix it ?

[firemonkey]

OK....Here are your results.

https://s5.postimg.org/bwkan233b/firemonkey2.png


My actual known ancestry (Known up to 3 gt level) - English 50% Scottish 31% Irish 19 %.

[MysteriousWays]

Could you please add me in this dimension as it seems more accurate? For example Albanians are plot north of Greeks as it should be on this dimension but on the other one we didn't

Sure. For everyone, the problem with PuntDNAL k15 is that it has 5 dimensions for Africa versus only three that cover Europe, so this would probably work very well for African populations, but is not that well oriented for European populations (MENA somewhere in between I think). And yes, here is your new graph...It still wants to pull you towards Italy.

https://s5.postimg.org/4j4wo3j1j/kelmendasi_v2.png

[Kelmendasi]

Sure. For everyone, the problem with PuntDNAL k15 is that it has 5 dimensions for Africa versus only three that cover Europe, so this would probably work very well for African populations, but is not that well oriented for European populations (MENA somewhere in between I think). And yes, here is your new graph...It still wants to pull you towards Italy.

https://s5.postimg.org/4j4wo3j1j/kelmendasi_v2.png
Thanks. Yh I have a western pull when compared to the average Albo although idk what component does this out of WHG, EEF and IE

[MysteriousWays]

thank you

I cant read much in the second map . any chance you can fix it ?

I can, but the results are a bit weird. The African dimensions are wreaking havoc on this.

https://s5.postimg.org/bavbqy813/All-_In3.png

[MysteriousWays]

I'm going to close this thread down and open up a new PCA with a different calculator. These African dimensions are just too problematic with how they are set up and it wreaks havoc on the components. Sorry about this.