PCA Plot of Greeks and neighbours (mostly)

[Danaan]

I made this PCA plot based on data from MDLP K23b (which I don't like).
I added a prediction of where I should plot. I haven't done any tests and I won't, until ancient greek and roman genomes get published.

I included the neighbors, the groups with which the Greeks or some of the Greeks have interacted and a few others.

http://i.imgur.com/DNKY9Sc.png

I hope people understand that towards French South there's more EEF or towards Ukrainians there's more EHG etc.

[Sikeliot]

This is based on MDLP K23, an outdated calculator. But it does show how most mainlanders do shift toward NE Euro Slavs.

[GoneWithTheWind]

Greeks plot all over the place

[Percivalle]

I hope people understand that towards French South there's more EEF or towards Ukrainians there's more EHG etc.

I hope you understand that French South means more WHG.

[Danaan]

I hope you understand that French South means more WHG.

How so? French South is the region with the highest EEF admixture after Basques and Sardinians.

[Percivalle]

How so? French South is the region with the highest EEF admixture after Basques and Sardinians.

French Southwest, not French South, and they still have high WHG. In this sense the taxonomy of MDLP is obsolete and outdated.

[Ibericus]

How so? French South is the region with the highest EEF admixture after Basques and Sardinians.
Basques have one of the highest levels of WHG,

[Danaan]

The data are from Gedmatch spreadsheet. The calculator has 53.58% Early European Farmers, 19,72% European Hunter Gatherers for 'French South'. French Basques get ~71% EEF.

Loschbour gets that result:
Population
Amerindian -
Ancestral_Altaic -
South_Central_Asian -
Arctic -
South_Indian -
Australoid -
Austronesian -
Caucasian -
Archaic_Human 0.06
East_African -
East_Siberian -
European_Early_Farmers 10.90
Khoisan -
Melano_Polynesian 0.50
Archaic_African 0.28
Near_East -
North_African -
Paleo_Siberian -
African_Pygmy -
South_East_Asian -
Subsaharian -
Tungus-Altaic -
European_Hunters_Gatherers 88.21

I don't know if the calculator is outdated, but I think that is correct.

[Percivalle]

I don't know if the calculator is outdated, but I think that is correct.

The name of the populations/components are outdated and obsolete, because now we know more about ancient genomes.

[Ibericus]

The data are from Gedmatch spreadsheet. The calculator has 53.58% Early European Farmers, 19,72% European Hunter Gatherers for 'French South'. French Basques get ~71% EEF.

Loschbour gets that result:
Population
Amerindian -
Ancestral_Altaic -
South_Central_Asian -
Arctic -
South_Indian -
Australoid -
Austronesian -
Caucasian -
Archaic_Human 0.06
East_African -
East_Siberian -
European_Early_Farmers 10.90
Khoisan -
Melano_Polynesian 0.50
Archaic_African 0.28
Near_East -
North_African -
Paleo_Siberian -
African_Pygmy -
South_East_Asian -
Subsaharian -
Tungus-Altaic -
European_Hunters_Gatherers 88.21

I don't know if the calculator is outdated, but I think that is correct.
The EEF component is outdated, because it has a lot of WHG hidden in it. It's a mix of WHG and Neolithic Levant. If you look at the latest Eurogenes test, the Basal-Rich K7, the Villabruna component in modern populations peaks in Baltic and Basques.

[Coolguy1]

Interesting how Smyrniot Greeks plot simmilar to island Greeks and not to Anatolian Turks. I think the northern component in mainland Greece is ancient, possible from the Dorian invasion and not really from the Slavs.

[Danaan]

Maybe. I don't trust Davidski though and in general I don't think that even 'experts' know much (they pretend to).
I used that calculator only because I had that data and I don't pretend that I have any special knowledge on the subject of genetics or statistics.

[Danaan]

Interesting how Smyrniot Greeks plot simmilar to island Greeks and not to Anatolian Turks. I think the northern component in mainland Greece is ancient, possible from the Dorian invasion and not really from the Slavs.

Smyrniot Greeks descended mostly from islanders who migrated there after the 17th century because the greek element had diminished there. My ancestor who is from there had a Cretan surname. That makes 100% sense.

[Coolguy1]

Smyrniot Greeks descended mostly from islanders who migrated there after the 17th century because the greek element had diminished there. My ancestor who is from there had a Cretan surname. That makes 100% sense.

Really? Im sure there were still Greek enclaves around the coast but it makes sense that islanders migrated to Anatolia. I wonder what the demographic status was like before the Turks arrived. I think its generally agreed upon that the majority of Anatolia was Greek speaking but im not sure how the people actually viewed themselves/referred upon themselves.

[Petalpusher]

I made this PCA plot based on data from MDLP K23b (which I don't like).
I added a prediction of where I should plot. I haven't done any tests and I won't, until ancient greek and roman genomes get published.

I included the neighbors, the groups with which the Greeks or some of the Greeks have interacted and a few others.

http://i.imgur.com/DNKY9Sc.png

I hope people understand that towards French South there's more EEF or towards Ukrainians there's more EHG etc.

No it's more WHG, here s the K7 peaks of Villabruna related admixture:


Villabruna 100%
Loschbour 93,04%
Motala 70,35%

Basque 62,66%
Lithuanian 60,31%
French South 59,70%
Estonian 59,69%
Latvian 59,59%
Orcadian 58,04%
Swedish 57,9%
German 56,88%
English 56,69%


Southern Euro:

Portuguese 52,36%
North Italian 51,5%
Serbian 50,45%
Tuscan 47,36%
Greece 45,08%
Abruzzo 40,74%
Sicilian 39,55%
Italian South 38,70%
Ashkenazi 35,73%
Cypriot 30,68%

Lebanese Druze 25,53%

[Danaan]

Older calculators had an EEF, Mediterranean etc component which peaked in Sardinians, Basques and a European Hunter Gatherer/NE Euro one which peaked in Baltic (like here) or in Finland (like Harappa World).
This is a PCA with that data. People can dismiss it.

I don't believe that Davidski's calculators were any good although that last one is interesting conceptually.

[Petalpusher]

Older calculators had an EEF, Mediterranean etc component which peaked in Sardinians, Basques and a European Hunter Gatherer/NE Euro one which peaked in Baltic (like here) or in Finland (like Harappa World).
This is a PCA with that data. People can dismiss it.

I don't believe that Davidski's calculators were any good although that last one is interesting conceptually.

That's why nobody is using it anymore, it hides too much HG for some people. You are probably misunderstanding what is early farmer, European admixture sure isn't in the direction of Lebanon... and Sicily does not have more indigenous European admixture than Lombards, it shouldn't take too much thoughts to figure this out. On the pca it's somewhere in the bottom right corner.

[safinator]

That's why nobody is using it anymore, it hides too much HG for some people. You are probably misunderstanding what is early farmer, European admixture sure isn't in the direction of Lebanon... and Sicily does not have more indigenous European admixture than Lombards, it shouldn't take too much thoughts to figure this out. On the pca it's somewhere in the bottom right corner.

Which Gedmatch calculator is the most up to date?

[Danaan]

That's why nobody is using it anymore, it hides too much HG for some people. You are probably misunderstanding what is early farmer, European admixture sure isn't in the direction of Lebanon... and Sicily does not have more indigenous European admixture than Lombards, it shouldn't take too much thoughts to figure this out. On the pca it's somewhere in the bottom right corner.

The pca doesn't have to correspond to geography.

[Danaan]

Do you like that more?
http://i.imgur.com/5q9F3Wu.png

[Petalpusher]

Which Gedmatch calculator is the most up to date?

For recent ancestry i d say K15 is still a good calculator, people usually find it accurate (when they don't have a crazy complicated ancestry). For deeper more modern autosomal based on adna, i guess the new K7, i like the Eurasia K10 and K6. Some mdlp are not bad (the old one are terrible), puntdnal as well.


The pca doesn't have to correspond to geography.
Im not sure what you are implying, why would it have to? genetic doesn't follow geography, yes, relatively. Greece for example is more northern than what the geography would suggest (basically C-S.Italian or something)


Do you like that more?

It's a bit more intuitive, but WHG is now in the upper left corner.

[brennus dux gallorum]

I made this PCA plot based on data from MDLP K23b (which I don't like).
I added a prediction of where I should plot. I haven't done any tests and I won't, until ancient greek and roman genomes get published.

I included the neighbors, the groups with which the Greeks or some of the Greeks have interacted and a few others.

http://i.imgur.com/DNKY9Sc.png

I hope people understand that towards French South there's more EEF or towards Ukrainians there's more EHG etc.

Which islands are those with so big distance from Athens, cause as far as I know cyclades and Ionian islands almost overlap pelloponese

Also is it possible for pelloponese to be closer to Macedonia than central Greece to Macedonia?

[Ibericus]

Btw where is the Italian_North sample from ? Im guessing they are from Friul or Veneto, because they are different from the Bergamos.

[Petalpusher]

Btw where is the Italian_North sample from ? Im guessing they are from Friul or Veneto, because they are different from the Bergamos.

It might be Friuli, i remember now mdlp k13 had one labelled Friuli so it must be that. It confirms it's actually the one we were talking about in another thread, it's not Piedmont, and the one i often get somewhere in single pop or mix mode.

[Danaan]

Which islands are those with so big distance from Athens, cause as far as I know cyclades and Ionian islands almost overlap pelloponese

Also is it possible for pelloponese to be closer to Macedonia than central Greece to Macedonia?

It has something to do with the sampling. Some studies don't include information on sampling methods. Peloponnese is diverse really so I don't think that the samples are representative of all the regions.

There's difference between Greek Macedonia and Greek Thessaloniki, for example. Apparently there is a significant number of people with Pontic Greek ancestry in the 'Greek Macedonia' sample. 'Greek islands' may be from Dodecanese (?). The distance isn't so big really. It appears big to some people because their nations are product of inbreeding.

In reality the most significant difference between Islands (some of them), Smyrna, Crete and the mainland is the lower 'European_Hunters_Gatherers'.

Before the Greeks there were Karians in the Aegean (maybe the Cycladic civilization was Karian?). I don't know if something can be attributed to them. But imo there are post-classical northern (NE and NW) influences in Greece, so I personally expect ancient Greeks to be around Greek Crete, Italia South, Greek Athens.

[Percivalle]

Btw where is the Italian_North sample from ? Im guessing they are from Friul or Veneto, because they are different from the Bergamos.


Yes, they might be North-eastern Italians, Veneto or Friuli.



Maybe. I don't trust Davidski though and in general I don't think that even 'experts' know much (they pretend to).
I used that calculator only because I had that data and I don't pretend that I have any special knowledge on the subject of genetics or statistics.


None of these amateur calculators is fully accurate or trustworthy. They all must be taken with a grain of salt. Even when these calculators use academic sample, often there is a bad habit to use only part of the sample, and this is clearly wrong.



It might be Friuli, i remember now mdlp k13 had one labelled Friuli so it must be that. It confirms it's actually the one we were talking about in another thread, it's not Piedmont, and the one i often get somewhere in single pop or mix mode.

MDLP 27Kb has two North Italian samples, K23b has three North Italian samples. Have you a link to the K23b spreadsheet?

If Italian_Piedmont is Val Borbera is not very representative of general Piedmontese population because it's considered a genetic isolate population and they are more southern-shifted than the average population.

[Sikeliot]

Which islands are those with so big distance from Athens, cause as far as I know cyclades and Ionian islands almost overlap pelloponese

Rhodes. It is a Rhodian sample.

[brennus dux gallorum]

Rhodes. It is a Rhodian sample.

OK

[Skerdilaid]

Interesting how Smyrniot Greeks plot simmilar to island Greeks and not to Anatolian Turks. I think the northern component in mainland Greece is ancient, possible from the Dorian invasion and not really from the Slavs.

Four Smyrna Greeks from this (http://bmcevolbiol.biomedcentral.com/articles/10.1186/1471-2148-11-69) study are actually paternally related to me, one of them really closely.

[Voskos]

Four Smyrna Greeks from this (http://bmcevolbiol.biomedcentral.com/articles/10.1186/1471-2148-11-69) study are actually paternally related to me, one of them really closely.

what does the TMRCA suggest? Arvanite origin or is it more ancient?

[Skerdilaid]

what does the TMRCA suggest? Arvanite origin or is it more ancient?

Most likely Arvanit and one of them even a Gheg, matching me 35/37.