PDA

View Full Version : Share Stanford Interpretome Results



Telfermagne
08-26-2013, 02:44 AM
Greetings, I am starting this thread for members to share their results from Stanford's Interpretome utility. I would like to compare how Interpretome stacks up against Dr. McDonald, 23andMe, etc. (mostly concerned with comparing against Dr. McDonald given the ISOGG comparisons between admix utilities - 23andMe = leader in commercial, Dr. McDonald offers a refinement over the leader in commercial).

My contribution (PCA for now using POPRES, still need to utilize the other features) -

When I ran Interpretome's PCA, using the POPRES dataset plotting PC 1 vs. PC 4 I got a picture that was consistent with Dr. McDonald's PCA. Interpretome yielded British with what seems to be a pull towards the South (& mayhaps a bit to the East as well); McDonald yielded close to the same but demonstrated a stronger Southward pull than Interpretome that put me closer to the French than British.

Screencaps (Interpretome vs. Dr. McDonald) -


V3:
http://i1161.photobucket.com/albums/q506/sar1227/InterpretomePCA1amp4POPRESV3Screenshot2013-08-25at100521PM_zpsbd1c057c.png (http://s1161.photobucket.com/user/sar1227/media/InterpretomePCA1amp4POPRESV3Screenshot2013-08-25at100521PM_zpsbd1c057c.png.html)

V2:
http://i1161.photobucket.com/albums/q506/sar1227/InterpretomePCA1amp4POPRESV2Screenshot2013-08-25at101811PM_zpsc2c91788.png (http://s1161.photobucket.com/user/sar1227/media/InterpretomePCA1amp4POPRESV2Screenshot2013-08-25at101811PM_zpsc2c91788.png.html)

McDonald:
http://i1161.photobucket.com/albums/q506/sar1227/Seth_Reeder_Full_20130209085122BGA3_zpseaf7e35b.pn g (http://s1161.photobucket.com/user/sar1227/media/Seth_Reeder_Full_20130209085122BGA3_zpseaf7e35b.pn g.html)

AJL
08-26-2013, 04:45 AM
Plotting x,y as axes 1,4, I am in Switzerland. On McDonald's map I am in the space between France, Spain, and Italy.

Sein
10-11-2013, 09:08 AM
My chromosome paintings:
Hapmap 2
756

Hapmap 3
757

HGDP World PCA:
763

Sapporo
10-11-2013, 09:23 AM
Sein, could you explain what the various world populations are on the HGDP World PCA?

Sein
10-11-2013, 10:56 AM
Razib Khan has left his genome out in the wild, so I tried these tools on his raw-data.

Hapmap 2
759

Hapmap 3
760

HGDP World PCA:
761

(My HGDP PCA for comparison)

Sein
10-11-2013, 11:02 AM
Oh, I'm sorry about that Sapporo, I posted before I read your question. The HGDP dataset includes Sub-Saharan Africans, North Africans-Southwest Asians, Europeans, Central-South Asians (South Asians are all Pakistani), East Asians, Oceanians, and Native Americans. I'm clustering with the HGDP Pakistanis. As always, the HGDP European, Middle Eastern, and Pakistani populations cluster very close on PCA (I guess that's why they pretty much define West Eurasia). I'm on the eastern edge of the HGDP Pashtuns, but still less East Eurasian shifted than the Burusho. Razib is clustering with the most West Eurasian shifted Hazara and Uyghur, which makes sense, because he tends to be construed as 55%-65% West Eurasian on various analyses.

Sein
10-11-2013, 11:13 AM
Also, to me, the most striking thing on PCA analyses that involve all of the continents is the intermediate nature of West Eurasians versus Sub-Saharan Africans and East Eurasians. The HGDP Middle Eastern, European, and Pakistani populations are really in between Sub-Saharan Africans and East Asians-Native Americans. Makes sense geographically, but people rarely look at it this way. Conceptualizing genetic structure as proceeding via tree-like differentiation is why people tend to say that the largest dimension of variance involves Sub-Saharan Africans versus all Eurasians, while the second largest dimension of variance involves West versus East Eurasians. But that doesn't make sense, since West Eurasians are significantly closer to Sub-Saharan Africans in comparison to East Eurasians, who show much greater distance to Sub-Saharan Africans. Also, I've been very redundant with the PCA plots, instead of posting two separate ones for myself and Razib, I could have just posted this one, which has both of us.

764

Sapporo
10-11-2013, 09:22 PM
Oh, I'm sorry about that Sapporo, I posted before I read your question. The HGDP dataset includes Sub-Saharan Africans, North Africans-Southwest Asians, Europeans, Central-South Asians (South Asians are all Pakistani), East Asians, Oceanians, and Native Americans. I'm clustering with the HGDP Pakistanis. As always, the HGDP European, Middle Eastern, and Pakistani populations cluster very close on PCA (I guess that's why they pretty much define West Eurasia). I'm on the eastern edge of the HGDP Pashtuns, but still less East Eurasian shifted than the Burusho. Razib is clustering with the most West Eurasian shifted Hazara and Uyghur, which makes sense, because he tends to be construed as 55%-65% West Eurasian on various analyses.

Interesting. What I find peculiar is that the HGDP Pakistani populations seem closer on PCA plots to far Eastern Europeans (I wonder if these are Russians?) than to North African/Southwest Asian populations. That is what 23andMe's World PCA plot shows as well. Although, I think the PCA plot lacks West Asians like Georgians, Armenians and Iranians. Still, do you feel that is accurate? That the HGDP Pakistanis are closer to far Eastern Europeans than to North Africans/Southwest Asians?


Also, to me, the most striking thing on PCA analyses that involve all of the continents is the intermediate nature of West Eurasians versus Sub-Saharan Africans and East Eurasians. The HGDP Middle Eastern, European, and Pakistani populations are really in between Sub-Saharan Africans and East Asians-Native Americans. Makes sense geographically, but people rarely look at it this way. Conceptualizing genetic structure as proceeding via tree-like differentiation is why people tend to say that the largest dimension of variance involves Sub-Saharan Africans versus all Eurasians, while the second largest dimension of variance involves West versus East Eurasians. But that doesn't make sense, since West Eurasians are significantly closer to Sub-Saharan Africans in comparison to East Eurasians, who show much greater distance to Sub-Saharan Africans. Also, I've been very redundant with the PCA plots, instead of posting two separate ones for myself and Razib, I could have just posted this one, which has both of us.

764

I've noticed this as well. In the case of North Africans and Southwest Asians, are they closer to Sub-Saharan Africans than to East Eurasians? It seems geography plays a large role here. West Eurasian populations closest to Sub-Saharan Africans geographically are closest to them genetically as well. Some of the North African/Southwest Asians have a pull in the direction of Sub-Saharan African while the HGDP Pakistani populations in the direction of East Eurasians.

Sein
10-11-2013, 10:40 PM
This is a very consistent finding throughout various analyses. Based on ChromoPainter/FineStructure, Europeans and the HGDP Pakistani populations always cluster together, to the exclusion of Southwest Asians-North Africans. One can see this in basic neighbor-joining trees as well. When analyzing West Eurasians, but not including West Asians and people from the Caucasus, Europeans and Pakistanis constitute a clade that excludes Southwest Asians. And European populations show much lower Fst distances to Pakistanis than to North Africans. Basically, Europeans and Pakistanis are much more genetically similar to each other than they are to populations like Palestinians, Jordanians, Bedouins, or Mozabites. West Asians and Caucasians make things more complicated, as Europeans and South Asians are closer to them than to each other. In fact, one can say that West Eurasians are split like this: Europeans-West Asians-Pakistanis-Northwest Indians vs Southwest Asians-North Africans. When one goes deeper, it's Europeans-West Asians vs Pakistanis-Northwest Indians, but West Asians have a distinct affinity to Pakistanis-Northwest Indians lacking amongst Europeans.

What you've stated is perfect, Southwest Asians-North Africans are Sub-Saharan African shifted, while Pakistanis-Northwest South Asians are East Eurasian shifted, and Europeans-West Asians are exactly intermediate between Sub-Saharan Africans and East Eurasians.

Sapporo
10-11-2013, 11:23 PM
This is a very consistent finding throughout various analyses. Based on ChromoPainter/FineStructure, Europeans and the HGDP Pakistani populations always cluster together, to the exclusion of Southwest Asians-North Africans. One can see this in basic neighbor-joining trees as well. When analyzing West Eurasians, but not including West Asians and people from the Caucasus, Europeans and Pakistanis constitute a clade that excludes Southwest Asians. And European populations show much lower Fst distances to Pakistanis than to North Africans. Basically, Europeans and Pakistanis are much more genetically similar to each other than they are to populations like Palestinians, Jordanians, Bedouins, or Mozabites. West Asians and Caucasians make things more complicated, as Europeans and South Asians are closer to them than to each other. In fact, one can say that West Eurasians are split like this: Europeans-West Asians-Pakistanis-Northwest Indians vs Southwest Asians-North Africans. When one goes deeper, it's Europeans-West Asians vs Pakistanis-Northwest Indians, but West Asians have a distinct affinity to Pakistanis-Northwest Indians lacking amongst Europeans.

What you've stated is perfect, Southwest Asians-North Africans are Sub-Saharan African shifted, while Pakistanis-Northwest South Asians are East Eurasian shifted, and Europeans-West Asians are exactly intermediate between Sub-Saharan Africans and East Eurasians.

Thank for your explanation. Overall, these findings make sense but I do feel we're ignoring the inter European variation. I presume Southern Europeans and HGDP North Africans/Southwest Asians have an affinity that Northern Europeans/Eastern Europeans lack but share to an extent with South-Central Asians/HGDP Pakistanis and West Asians/Caucasians of course. So, while Europeans are of course closest to one another, they share different affinities with other world populations.

Based on this data, I just find it very interesting that a HGDP Pashtun or Sindhi is much closer to a Russian or Finn genetically than they are to a Palestinian, Egyptian or Moroccan. MENA is nowhere as unified as many people like to presume.

Sapporo
10-11-2013, 11:37 PM
MENA is a geopolitical term...

Yes, I realize that. However, that doesn't stop people from assuming MENA are all close to one another genetically.

Sein
10-11-2013, 11:59 PM
There are some nice papers dealing with this topic, fascinating stuff:
http://www.plosgenetics.org/article/info%3Adoi%2F10.1371%2Fjournal.pgen.1003447
http://arxiv.org/pdf/1303.7475v1.pdf (no tree, but they find that Pakistani and European populations cluster together, to the exclusion of Southwest Asians)
http://www.plosone.org/article/info:doi/10.1371/journal.pone.0049438
http://www.sciencemag.org/content/319/5866/1100.full.pdf
http://www.plosgenetics.org/article/info%3Adoi%2F10.1371%2Fjournal.pgen.1002453
http://www.plosgenetics.org/article/info%3Adoi%2F10.1371%2Fjournal.pgen.1003316 (In this paper, "Central Asian" means HGDP Pakistani. They do note, "Consequently, Levant populations today fall into two main groups: one sharing more genetic characteristics with modern-day Europeans and Central Asians, and the other with closer genetic affinities to other Middle Easterners and Africans.")

I have a tentative suggestion about all of this. Perhaps the African ancestry in Southwest Asians and North Africans is the reason for the greater divergence between them and other West Eurasians. In addition, European and (based on what Reich told me) West Asian populations have substantial East Eurasian-American admixture, which in a general sense resembles the East Eurasian admixture (or, if you prefer, "South Eurasian" admixture) in the HGDP Pakistani samples. Not quite sure, but I think this is plausible. In fact, Dienekes' once posted an experiment in which he found that the more "ASI" a South Asian population was, the greater the D-statistics between it and Russians. Pretty cool and unexpected. Also, two neighbor joining trees that illustrate all of this:

768

769

As expected, Europeans and HGDP Pakistanis cluster together, versus Middle Easterners.

Sapporo
10-12-2013, 12:46 AM
(In this paper, "Central Asian" means HGDP Pakistani. They do note, "Consequently, Levant populations today fall into two main groups: one sharing more genetic characteristics with modern-day Europeans and Central Asians, and the other with closer genetic affinities to other Middle Easterners and Africans.")

Thank you for the papers. I will go through them later. I think part of the difference in the Levant is between the Northern and Southern Levant. There may also be differences in admixture due to religious reasons.



I have a tentative suggestion about all of this. Perhaps the African ancestry in Southwest Asians and North Africans is the reason for the greater divergence between them and other West Eurasians. In addition, European and (based on what Reich told me) West Asian populations have substantial East Eurasian-American admixture, which in a general sense resembles the East Eurasian admixture (or, if you prefer, "South Eurasian" admixture) in the HGDP Pakistani samples. Not quite sure, but I think this is plausible. In fact, Dienekes' once posted an experiment in which he found that the more "ASI" a South Asian population was, the greater the D-statistics between it and Russians. Pretty cool and unexpected. Also, two neighbor joining trees that illustrates all of this:

768

769

As expected, Europeans and HGDP Pakistanis cluster together, versus Middle Easterners.



I think your suggestion is quite credible. Especially, in the case of Mozabites and to a lesser extent Egyptians. Although, I believe the HGDP Bedouin and Palestinian samples don't have much African admixture. Your point about a shared East-Eurasian/American like admixture in West Asians and Northern Europeans also makes sense as the HGDP Pakistanis are closer to both West Asians and Northeast Europeans than to North Africans or Southwest Asians. I'd include Southern Europeans as well.

Although, in regards to the Druze, on the neighbor-joining tree, are they closer to West Asians and Northern Europeans than the HGDP Pakistanis?

Sein
10-12-2013, 01:07 AM
Surprisingly, the Druze are clustering with other Middle Easterners in these two trees, rather than with Europeans and Pakistanis. Depending on the paper, the Druze cluster either with Middle Easterners, or with Europeans-Pakistanis. It seems their case is more complicated.

Sapporo
10-12-2013, 01:09 AM
Surprisingly, the Druze are clustering with other Middle Easterners in these two trees, rather than with Europeans and Pakistanis. Depending on the paper, the Druze cluster either with Middle Easterners, or with Europeans-Pakistanis. It seems their case is more complicated.

Yes, it seems like they have an affinity to both populations. I think it can be seen on McDonald plots as well.

indigo
02-17-2014, 09:43 AM
Sorry if resurrecting an older thread, but where do we find the information of which populations group together? For example, who falls under CHB+JPT on Hapmap 2 (for example, do Middle Easterners group together with the CHB+JPT or CEU)?
Also, I am curious if there is any one setting on the Interpretome chromosome painting tool that yields the most accurate results?
And lastly… Can anyone make their Interpretome chromosome painting show results from all populations (or I suppose vice-versa, only one population) depending on the settings?

Ebizur
02-17-2014, 01:52 PM
Thank you for the papers. I will go through them later. I think part of the difference in the Levant is between the Northern and Southern Levant. There may also be differences in admixture due to religious reasons.

I think your suggestion is quite credible. Especially, in the case of Mozabites and to a lesser extent Egyptians. Although, I believe the HGDP Bedouin and Palestinian samples don't have much African admixture. ...

Although, in regards to the Druze, on the neighbor-joining tree, are they closer to West Asians and Northern Europeans than the HGDP Pakistanis?

The HGDP Bedouin, Palestinian, and Mozabite samples seem to contain a lot of obvious African admixture:

HGDP Bedouin Y-DNA (n=27)
1/27 = 0.037 B

1/27 = 0.037 E1b1b1a3
1/27 = 0.037 E1b1b1b (basal member of "Berber" group)
2/27 = 0.074 E1b1b1c
4/27 = 0.148 E1b1b1 total

1/27 = 0.037 G (MRCA with a Pathan G approx. 6,000 YBP; MRCA with a Russian/Adyghe branch of G approx. 16,000 YBP)
19/27 = 0.704 J(xJ2)
2/27 = 0.074 R1a1

HGDP Bedouin mtDNA (n=27)
1 x U7: MRCA Hazara (U7)/approx. 4,000 YBP
1 x L0a: MRCA Makrani (L0a)/approx. 6,000 YBP
1 x K1a: MRCA Druze (K1a)/approx. 6,000 YBP
1 x H2a: MRCA Adyghe (H2a)/approx. 6,500 YBP
1 x L2a: MRCA Makrani (L2a)/approx. 7,500 YBP
1 x J1c: MRCA Oroqen (J1c)/approx. 7,500 YBP
1 x L1b: MRCA {Mandenka + Mandenka} (L1b)/approx. 8,500 YBP
2 x R0a: MRCA {Palestinian + Pathan} (R0a)/approx. 9,000 YBP
1 x HV: MRCA Hazara (HV)/approx. 9,000 YBP
1 x H1: MRCA {Sardinian + {Basque + {Basque + Basque}}} (H1/H1t)/approx. 9,500 YBP
1 x L0a: MRCA Bantu (L0a)/approx. 11,500 YBP
1 x N1b: MRCA Palestinian (N1b)/approx. 12,500 YBP
1 x T1a: MRCA {{French + Pathan} + {Basque + {Pathan + {Brahui + Brahui}}}} (T1a)/approx. 13,500 YBP
1 x H: MRCA Makrani (H14)/approx. 14,000 YBP
1 x L2a: MRCA {Mandenka + Mandenka} (L2a)/approx. 14,000 YBP
1 x K1a: MRCA {{Adyghe + {Basque + French}} + {Druze + Bedouin}} (K1a)/approx. 14,500 YBP
1 x HV1: MRCA Palestinian (HV1)/approx. 15,000 YBP
1 x X2e: MRCA {Makrani + {Burusho + Burusho}} (X2)/approx. 16,000 YBP
1 x I5a: MRCA Palestinian (I)/approx. 17,500 YBP
1 x L0a: MRCA {Bantu + {Bantu + Bedouin}} (L0a)/approx. 18,000 YBP
1 x J2a: MRCA {Baloch + {Mozabite + Mozabite}} (J2a)/approx. 18,000 YBP
1 x HV: MRCA Hazara (HV)/approx. 18,500 YBP
1 x R0a: MRCA {{Druze + {{Palestinian + Pathan} + {Bedouin + Bedouin}}} + {Kalash + {Kalash + {Kalash + Kalash}}}} (R0a)/approx. 26,500 YBP
1 x N1a: MRCA all {N1e + I}/approx. 37,000 YBP
1 x M1a: MRCA rest of M/approx. 61,000 YBP
1 x L2b: MRCA {Mandenka + {Mandenka + Mandenka}} (L2c)/approx. 62,500 YBP

1/27 = 0.037 B, 7/27 = 0.259 L(xM, N), 8/54 = 0.148 "Sub-Saharan African" haplogroup total

HGDP Palestinian Y-DNA (n=16)
2/16 E1b1b1a3
13/16 G (11/16 coalesce to a MRCA approx. 2,000 YBP; this group coalesces with a broad group of G Y-chromosomes from Italy (including Sardinia), Adyghe, Brahui, and a Pima approx. 10,000 YBP)
1/16 J(xJ2)

HGDP Palestinian mtDNA (n=16)
1 x M30: MRCA Sindhi (M30)/approx. 2,000 YBP
1 x U5a: MRCA Hazara (U5a)/approx. 4,500 YBP
1 x H: MRCA {Mozabite + Orcadian} (H/H23)/approx. 5,500 YBP
1 x HVO: MRCA Basque (HVO)/approx. 7,000 YBP
1 x T2b: MRCA {Burusho + {Russian + Russian}} (T2b)/approx. 7,000 YBP
1 x R0a: MRCA Pathan (R0a)/approx. 7,500 YBP
1 x U1a: MRCA Adyghe (U1a)/approx. 8,500 YBP
1 x H13: MRCA Tuscan (H13)/approx. 9,000 YBP
1 x L2a: MRCA {Bedouin + Makrani} (L2a)/approx. 10,500 YBP
1 x N1b: MRCA Bedouin (N1b)/approx. 12,500 YBP
1 x L3e: MRCA Yoruba (L3e)/approx. 13,000 YBP
1 x U1a: MRCA Burusho (U1a)/approx. 14,500 YBP
1 x HV1: MRCA Bedouin (HV1)/approx. 15,000 YBP
1 x HV1: MRCA {Palestinian + Bedouin} (HV1)/approx. 17,000 YBP
1 x I: MRCA Bedouin (I5a)/approx. 17,500 YBP
1 x U6a: MRCA {Mozabite + {Mozabite + Mozabite}} (U6a)/approx. 23,000 YBP

2/16 = 0.125 L(xM, N), 2/32 = 0.063 "Sub-Saharan African" haplogroup total

HGDP Mozabite Y-DNA (n=20)
2/20 E1b1a
16/20 E1b1b1b
2/20 R1b1(xR1b1b1, R1b1b2)

HGDP Mozabite mtDNA (n=20)
1 x H: MRCA Orcadian (H23)/approx. 3,000 YBP
1 x V: MRCA Bergamo Italian (V3)/approx. 6,500 YBP
1 x H5a: MRCA {Russian + Hazara} (H5a)/approx. 8,000 YBP
1 x L3e: MRCA Bantu (L3e)/approx. 8,500 YBP
3 x H1v: MRCA {French + Druze} (H1u)/approx. 8,500 YBP
1 x V: MRCA {Mozabite + Bergamo Italian} (V/V3)/approx. 9,000 YBP
1 x L2a: MRCA {Palestinian + {Bedouin + Makrani}} (L2a)/approx. 12,000 YBP
1 x H: MRCA {Bergamo Italian + {Makrani + {{Bergamo Italian + {Hazara + Druze}} + {Palestinian + {Mozabite + Orcadian}}}}} (H/H7/H23)/approx. 12,000 YBP
2 x J2a: MRCA Baloch (J2a)/approx. 16,500 YBP
1 x L3e: MRCA {Mozabite + Bantu} (L3e)/approx. 22,500 YBP
3 x U6a: MRCA Palestinian (U6a)/approx. 23,000 YBP
1 x U6a: MRCA {Palestinian + {Mozabite + {Mozabite + Mozabite}}} (U6a)/approx. 26,000 YBP
1 x U8b: MRCA K/approx. 33,000 YBP
2 x U3a: MRCA U4'7'8'9/approx. 47,000 YBP

2/20 = 0.100 E1b1a, 3/20 = 0.150 L(xM, N), 5/40 = 0.125 "Sub-Saharan African" haplogroup total

On the other hand, the HGDP Druze sample does not contain any haploid evidence of Sub-Saharan African admixture:

HGDP Druze Y-DNA (n=12)
1/12 E1b1b1a2
5/12 L
1/12 J(xJ2)
4/12 J2
1/12 R1b1b2

HGDP Druze mtDNA (n=12):
1 x H7b: MRCA Hazara (H7b)/approx. 2,500 YBP
1 x H1u: MRCA French (H1u)/approx. 6,000 YBP
1 x K1a: MRCA Bedouin (K1a)/approx. 6,000 YBP
2 x H1b: MRCA {Russian + Russian} (H1)/approx. 7,000 YBP
1 x U1a: MRCA Hazara (U1a)/approx. 8,500 YBP
1 x H4b: MRCA Orcadian (H4a)/approx. 11,500 YBP
1 x R0a: MRCA {{Palestinian + Pathan} + {Bedouin + Bedouin}} (R0a)/approx. 12,500 YBP
1 x R5a: MRCA Burusho (R5a)/approx. 13,500 YBP
1 x H26: MRCA H3 + H1/approx. 16,500 YBP
2 x H33: MRCA Brahui + Uyghur + Uyghur (H6a/H/H8)/approx. 16,500 YBP

Sein
02-17-2014, 01:57 PM
Sorry if resurrecting an older thread, but where do we find the information of which populations group together? For example, who falls under CHB+JPT on Hapmap 2 (for example, do Middle Easterners group together with the CHB+JPT or CEU)?
Also, I am curious if there is any one setting on the Interpretome chromosome painting tool that yields the most accurate results?
And lastly… Can anyone make their Interpretome chromosome painting show results from all populations (or I suppose vice-versa, only one population) depending on the settings?

CHB+JPT are Chinese and Japanese, CEU are White Americans from Utah. Middle Easterners will almost always cluster near CEU, but some might be shifted in an ASW direction. ASW are African Americans, and some Middle Easterners (especially Arabic speakers) and almost all North Africans have minor-moderate Sub-Saharan African admixture.

The best settings for HapMap 2:

1427

DMXX
03-06-2014, 01:23 AM
In fact, one can say that West Eurasians are split like this: Europeans-West Asians-Pakistanis-Northwest Indians vs Southwest Asians-North Africans. When one goes deeper, it's Europeans-West Asians vs Pakistanis-Northwest Indians, but West Asians have a distinct affinity to Pakistanis-Northwest Indians lacking amongst Europeans.



Based on this data, I just find it very interesting that a HGDP Pashtun or Sindhi is much closer to a Russian or Finn genetically than they are to a Palestinian, Egyptian or Moroccan. MENA is nowhere as unified as many people like to presume.

My results (HGDP World 100k SNP PC1x2 settings) are in accordance with the above. Note that I appear intermediate between the HGDP Pakistanis and Europeans (shifted more toward the latter?) rather than clustering with the ethnicities which constitute the Near Eastern cluster.

There may be a slight shift towards the Near Eastern cluster, although I still sit within the European-South Asian cline regardless.

[Edit]: Just inserted a zoomed in version. It's clear now I am firmly within the Euro-SC Asian cline. Also removed "Eastern European" from this message. I have no certainty which European group I'm clustering near.


http://i60.tinypic.com/rutnrd.png

http://i60.tinypic.com/2bv5vo.png

Humanist
03-06-2014, 04:17 AM
My results (HGDP World 100k SNP PC1x2 settings) are in accordance with the above. Note that I appear intermediate between the HGDP Pakistanis and Eastern Europeans (shifted more toward the latter?) rather than clustering with the ethnicities which constitute the Near Eastern cluster.

Doing the same for myself.

http://i1096.photobucket.com/albums/g326/dok101/Faces/mds_intepretome_1.jpg

http://i1096.photobucket.com/albums/g326/dok101/Faces/mds_intepretome_2.jpg

Sein
03-06-2014, 04:23 AM
Very interesting results. So DMXX does seem to be clustering in between South Asia and Eastern Europe, while you are clustering in between the Levant and Southern Europe.

ZephyrousMandaru
03-06-2014, 05:23 AM
Mine.

http://i26.photobucket.com/albums/c131/Darth_Azasyahigor/HGDP-WorldPC1and2.png?t=1394083215

http://i26.photobucket.com/albums/c131/Darth_Azasyahigor/HGDP-WorldPC1and2Zoomed.png?t=1394083239

vettor
03-06-2014, 09:50 AM
to all of the above, how accurate is your regional plots showing .?

mine is accurate as ancestors are north-italian

http://imageshack.com/scaled/medium/27/kul8.png (http://imageshack.com/photo/my-images/27/kul8.png/) Uploaded with ImageShack.us (http://imageshack.com)

instructions are to use Pca1 and Pca4

unsure if there are other combinations for different zones of the planet

newtoboard
03-06-2014, 01:17 PM
In fact, one can say that West Eurasians are split like this: Europeans-West Asians-Pakistanis-Northwest Indians vs Southwest Asians-North Africans. When one goes deeper, it's Europeans-West Asians vs Pakistanis-Northwest Indians, but West Asians have a distinct affinity to Pakistanis-Northwest Indians lacking amongst Europeans.


That is strange. North Africans and SW Asians are definitely more West Eurasian than NW South Asians. And SSA can't be the reason why this split occurs because ASI should pull NW South Asians just as far away (if not more) from Europeans and West Asians as SSA does for North Africans and SW Asians. Perhaps this has to do with what David coined as a split between Northwest Eurasians (represented by the various West Asian and Northern European components and perhaps some Mediterranean components as well) and Southwest Eurasians (represented by the various SW Asian, North African and Mediterranean centered components). The West Eurasian ancestry of NW South Asians is almost entirely related to NW Eurasians while the West Eurasian ancestry of North Africans and SW Asians is almost entirely related to SW Eurasians. Obviously most West Asians and Europeans have more ancestry from the Northwest Eurasian category and the ones that don't (like Greeks especially the ones lacking in the Northern European component (Cretans?)) seem to be some extra affinity to the Levant and Cyprus. The split between Europeans-West Asians and NW South Asians is probably the result of ASI and the West Asian affinity to NW South Asian populations lacking among Europeans is related to more of the West Asian component among West Asians and NW South Asians in comparison to the Northern European component.

newtoboard
03-06-2014, 01:20 PM
Very interesting results. So DMXX does seem to be clustering in between South Asia and Eastern Europe, while you are clustering in between the Levant and Southern Europe.

In a way could we say in terms of clustering Iranians are more of a Central Asian population? They probably represent what the Central Asian population prior to Turkification looked like better than most modern day Central Asian populations.

Sein
03-06-2014, 04:24 PM
I think ANE ancestry is what makes Northwestern South Asians and South Central Asians seem less West Eurasian. ANE seems to be much more "Eastern" shifted, in comparison to proper West Eurasian ancestry, because it lacks any connection with the African-like "Basal Eurasian" element throughout Europe and West Asia. But in terms of "clade" or "linage", Ancestral North Eurasians were in the same grouping of peoples as West Eurasians proper (WHG). In fact, you could say ANE is basically West Eurasian, but stripped of any African affinity. Davidski just did an ANE test for West, Central, and South Asians. I turned out around 36% ANE! This is much greater than any other population in Eurasia. Also, he found my East Eurasian admixture to be around 16%. In his run, the average West Asian/Caucasian/Southeast European is 6% East Eurasian. So, I'm just 10% more East Eurasian than the average West Asian/Caucasian/Southeast European. But my ANE percentage took me by surprise. I think this is very cool, to say the least. Now, I can claim to be ancestral to Native Americans and Europeans! B) ;)

(A side note, but this validates Geno 2.0 in a very nice manner. According to Geno 2.0, I am 82% West Eurasian, and 2% Native American. Add that 2% Native American to my West Eurasian estimate, and you end up with 84% West Eurasian. Also, I guess this means I'm 11% ASI, 4% Siberian. So, Geno 2.0 did turn out rather accurate)

I'm just joking/playing around here in regards to me being ancestral to Native Americans and Europeans, but on a more serious note, I wonder how ANE got to be so high in South Central Asia? I think DMXX's thoughts in this regard are highly pertinent, and deeply insightful. He shared some thoughts on this very subject, in this thread, http://www.anthrogenica.com/showthread.php?2200-History-of-Jatts-and-the-case-of-the-elevated-NE-Euro-comonent/page2, and after looking at my results, I think he has hit the nail right on the head.

MfA
03-06-2014, 04:37 PM
HDGP: World

http://abload.de/img/1nvqh6.png

HDGP: European (Detailed Region)

http://abload.de/img/2agu7i.png



Hapmap: World

http://abload.de/img/3dfopj.png

Hapmap: World (Local Clusters)

http://abload.de/img/4ogi30.png



POPRES: European (PCA1-4)

http://abload.de/img/5i2uk0.png



Mid-East/Jewish

http://abload.de/img/8ylude.png

Mid-East/Jewish (Local Clusters)

http://abload.de/img/9quupr.png



Referance Populations
PCA1-2

http://abload.de/img/7zvohl.png

PCA1-4

http://abload.de/img/6mdqc0.png

soulblighter
03-06-2014, 07:24 PM
I think ANE ancestry is what makes Northwestern South Asians and South Central Asians seem less West Eurasian. ANE seems to be much more "Eastern" shifted, in comparison to proper West Eurasian ancestry, because it lacks any connection with the African-like "Basal Eurasian" element throughout Europe and West Asia. But in terms of "clade" or "linage", Ancestral North Eurasians were in the same grouping of peoples as West Eurasians proper (WHG). In fact, you could say ANE is basically West Eurasian, but stripped of any African affinity. Davidski just did an ANE test for West, Central, and South Asians. I turned out around 36% ANE! This is much greater than any other population in Eurasia. Also, he found my East Eurasian admixture to be around 16%. In his run, the average West Asian/Caucasian/Southeast European is 6% East Eurasian. So, I'm just 10% more East Eurasian than the average West Asian/Caucasian/Southeast European. But my ANE percentage took me by surprise. I think this is very cool, to say the least. Now, I can claim to be ancestral to Native Americans and Europeans! B) ;)

I'm just joking/playing around here in regards to me being ancestral to Native Americans and Europeans, but on a more serious note, I wonder how ANE got to be so high in South Central Asia? I think DMXX's thoughts in this regard are highly pertinent, and deeply insightful. He shared some thoughts on this very subject, in this thread, http://www.anthrogenica.com/showthread.php?2200-History-of-Jatts-and-the-case-of-the-elevated-NE-Euro-comonent/page2, and after looking at my results, I think he has hit the nail right on the head.

Did David put up the results in a spreadsheet?

Here are my results

PCA1 vs PCA 2:
1555

PCA1 vs PCA 3:
1556

everest59
03-06-2014, 07:31 PM
Sein, that number is inaccurate. You will have between 10-20% ANE. What happens is, ANE likes to grab on to ENA quite a bit. The best way to calculate ANE for anybody with any amount of ENA is to rely on shared genetic drift with Mal'ta. I posted shared drift numbers a while ago. Pashtuns for example have lower shared drift number compared to Orcadians. The highest shared drift ratio was in Karitiana, which makes sense obviously.
Nonetheless, Pashtuns do have elevated ANE. So it will affect your positioning in a PCA.

Sein
03-06-2014, 07:39 PM
Here is what he's posted so far:
https://docs.google.com/spreadsheet/ccc?key=0Ato3EYTdM8lQdGE5bjNCdl8yci0yNGN4TWk4M3NUW VE&usp=sharing

He is going to add more populations, but he said it'll be a time-consuming process. Also, he told me about my particular results via personal communication. I'm pretty different from the Afghan Pashtun. The Afghan Pashtun has less ANE admixture, and more East Eurasian admixture, in comparison to myself.

Please do compare these results with the values found in the actual paper:
http://s129.photobucket.com/user/dpwes/media/Ancient_human_genomes_suggest_three_ancestral_popu lations_TableS1212.png.html

An almost perfect correspondence! It seems very accurate. Since it seems to fit very nicely with the paper, I think my percentage of ANE is probably correct. Although, I was pretty surprised that I turned out 36% ANE.

Edit: Everest makes a good point, this is all very interesting. Although, West and South Asians don't have any WHG admixture. So, Orcadians will share higher drift, since my EEF ancestry makes me more divergent from ANE, despite the high amount of shared ancestry.

everest59
03-06-2014, 08:50 PM
Well, if his method shows the Paniya or Dusadh have virtually no ANE then I'll trust his methodology. All the numbers look good though. The Pashtuns seem off.

Dr_McNinja
03-07-2014, 12:32 AM
^ On a semi-related note, I found another Punjabi (either Jatt or Rajput) who had Q1 Y-DNA haplogroup today on 23andMe.

Sein
03-07-2014, 02:50 AM
Well, if his method shows the Paniya or Dusadh have virtually no ANE then I'll trust his methodology. All the numbers look good though. The Pashtuns seem off.

The numbers do seem very accurate. And he was even able to separate ANE from ENA. If it is working for the Caucasians, West Asians, and Southeastern Europeans, I think it is probably working for the Pashtuns. And it matches rather nicely with my Geno 2.0 results. I am 82% West Eurasian according to Geno 2.0, and 2% Native American. If you add that 2% Native American into my West Eurasian, since it is probably ANE rather than ENA, I turn out 84% West Eurasian. And based on his methodology, I am 84% West Eurasian! So, the correspondence is pretty solid. Also, I think I can now divide my ENA ancestry up. I am probably 11% ASI, and 4% Siberian/Northeast Asian. This is also what Geno 2.0 had me at, 11% SE Asian (ASI) and 4% NE Asian. Also, I think another fact validates his method. The Afghan Pashtun sample has much less ANE than myself. If you look at the Eurogenes K13 results, they have a substantial European percentage stemming from the Mediterranean-Atlantic. By contrast, I seem to have no genetic connection to Southern Europe. And on PCA plots, this sample clusters closer to East Asians in comparison to myself, so it makes sense that I'm less ENA than them. In short, the results can't really be flat out wrong. In broad terms, they seem robust.

Also, most of my ancestry is Gedrosia/Baloch+Caucasian+NE Euro. So I think a high ANE percentage was always probable. But I must admit, 36% did catch me off guard. If correct, this is probably the highest ANE percentage yet found for all living Eurasian populations. But I can't be alone. I think all Pashtuns and Jatts will have huge ANE percentages, just like myself. In fact, perhaps I'm less ANE than the average Pashtun or Jatt! We'll find out when he adds more results. He has to run each sample individually, so it'll probably take some time.

Finally, I tried the ANE-WHG-EEF test for West Asians and Southeastern Europeans. I turned out 35% ANE! So, this does seem to be correct. But I absolutely agree, the real test of this are tribal South Indian people. If they show ANE, something is probably wrong. I suppose we just have to wait. Pretty excited about this, I wonder if his method will work with South Indians?

DMXX
03-07-2014, 09:20 AM
See this new thread (http://www.anthrogenica.com/showthread.php?2281-Anthrogenica-community-Interpretome-chart&p=32977#post32977) for something a bit more community-driven with respect to Interpretome's maps. I was going to include MfA as well, but he hasn't zoomed on his HGDP:World upload.

MfA
03-15-2014, 05:32 PM
parasar, but you just provided the evidence, no? There is no reason why I should be more similar to Greeks and Europeans (e.g. English, French) than Turks, unless Anatolian Turks have absorbed various "eastern" elements over the last 1000+ years (Turkic, Kurdish, etc.).

http://i1096.photobucket.com/albums/g326/dok101/Faces/popres_europe_hum.png

http://www.anthrogenica.com/attachment.php?attachmentid=1571&d=1394287552

That's certainly not Kurdish ancestry pulling them eastward otherwise I wouldn't be both western and northern compared to you..

http://abload.de/img/popres_mfa_xojo0.png

Humanist
03-15-2014, 05:38 PM
That's certainly not Kurdish ancestry pulling them eastward otherwise I wouldn't be both western and northern compared to you..

http://abload.de/img/popres_mfa_xojo0.png

Well, then, we will have to agree to disagree.

EDIT:

Based on the results of the below IBD heat map, I would have to agree with you regarding the Kurdish input in Turks. It does not register below. However, I still do not believe that, overall, your genome is most similar to W Europeans. I do agree that, overall, you may be more similar to E Europeans, and perhaps Scandinavians, however.

http://i1096.photobucket.com/albums/g326/dok101/Faces/heatmap.png

MfA
03-15-2014, 06:26 PM
Well, then, we will have to agree to disagree.

EDIT:

Based on the results of the below IBD heat map, I would have to agree with you regarding the Kurdish input in Turks. It does not register below. However, I still do not believe that, overall, your genome is most similar to W Europeans. I do agree that, overall, you may be more similar to E Europeans, and perhaps Scandinavians, however.


You are using Dodecad samples which is oversampled with Southern and Eastern Kurds not Northern/Western Kurds that one would expect the Kurdish contribution to Turks.

Neither do I. I'm using your argument, If you think you are western compared to Turks because you don't have those "eastern" elements which is basically East Eurasian in this case, I and overall Kurds don't have either since I'm western and northern compared to you. Therefore the reason Turks are eastern not because of Kurdish ancestry which is similar to other indigenous peoples like Armenians and Assyrians like yourself, but East Eurasian ancestry.

Humanist
03-15-2014, 06:40 PM
Therefore the reason Turks are eastern not because of Kurdish ancestry...

I have already conceded that fact.


Kurdish ancestry which is similar to other indigenous peoples like Armenians and Assyrians like yourself...

Kurdish ancestry is similar to Armenians and Assyrians. Yes. However, Kurds are most similar to Iranians. And, Iranians are rather distinct from Assyrians and Armenians. They are more South-Central Asian ("eastern") than either Armenians or Assyrians.

DMXX's Vaedhya blog (http://vaedhya.blogspot.com/2012/08/introducing-acd-tool.html):


Example #2: West Asians (Harappa)
Using the Harappa Ancestry Project this time, I ran the data of Armenians, Assyrians, Kurds and Iranians (mostly from the Harappa cohort) into the ACD Tool once more and presented the differences as above:

http://4.bp.blogspot.com/-TI1V30XXDw4/UDDQk1Dj11I/AAAAAAAAAOk/_LZxYUxQgso/s640/Example2.png

MfA
03-15-2014, 09:32 PM
I have already conceded that fact.
Then Why did you bring up Kurds when Kurds are not responsible of Turk's easterness?



Kurdish ancestry is similar to Armenians and Assyrians. Yes. However, Kurds are most similar to Iranians. And, Iranians are rather distinct from Assyrians and Armenians. They are more South-Central Asian ("eastern") than either Armenians or Assyrians.

DMXX's Vaedhya blog (http://vaedhya.blogspot.com/2012/08/introducing-acd-tool.html):

Irrelevant, who is close to Kurds or not, it's not the subject of this discussion.

Humanist
03-15-2014, 10:08 PM
Then Why did you bring up Kurds when Kurds are not responsible of Turk's easterness?

I made a mistake regarding the IBD data, and that is why I stated that "I have already conceded that fact."


Irrelevant, who is close to Kurds or not, it's not the subject of this discussion.

OK. We seem to be coming at this from two different perspectives. Let us leave it at that.

tippy
06-08-2016, 02:18 PM
I'm reviving this because I have a keen interest in seeing other Near Eastern/West Asian/North African Hapmap world results especially

hapmap 100k

me

http://i.imgur.com/NwpfaFu.png


mom


http://imgur.com/qhUOE0s.png


HGDP World (100k)

me

http://imgur.com/aEfQZxD.png

mom

http://imgur.com/kICASuE.png

MidEast/Jewish

me

http://imgur.com/e5qhMb6.png

mom

http://imgur.com/IVWTJ5z.png

Ok, my question for the learned. Why does hapmap world seem to show a lot more of an African drag than anywhere else? Also does anyone have any other Near Eastern hapmap results as I'd really be interested to see where everyone else lands.

firemonkey
06-08-2016, 04:06 PM
My HGDP (100/1-2) World and European. The European is far more accurate.

9689

9691


My Pop res European.

9692