Genetic and Cultural Differences between Jews and Greeks

[Riverman]

@Jovialis: Is ITA_Rome_Imperial:RMPR835 = R835_Imperial_Era_Civitanova_Marche?

Or is this wrong?

Because if I run this sample without Imperial Roman as a reference, I get:

0.05831675	GRC_Mycenaean
0.08316841	ITA_Rome_Latini_IA
0.08433978	Levant_LBN_Roman
0.09021476	TUR_Alalakh_MLBA
0.12287859	CZE_Hallstatt_Bylany
0.13108274	DEU_MA
0.13183519	CZE_Early_Slav
0.17643994	Berber_Tunisia_Chen
0.23509565	TUR_Ottoman
0.60633450	Dai
0.74575043	Yoruba

42.4	TUR_Alalakh_MLBA
26.6	ITA_Rome_Latini_IA
14.6	CZE_Early_Slav
6.2	DEU_MA
5.6	Berber_Tunisia_Chen
4.0	GRC_Mycenaean
0.4	Yoruba
0.2	Dai

If I add Imperial Roman:

Distance to:	ITA_Rome_Imperial:RMPR835
0.03676210	ITA_Rome_Imperial
0.05831675	GRC_Mycenaean
0.08316841	ITA_Rome_Latini_IA
0.08433978	Levant_LBN_Roman
0.09021476	TUR_Alalakh_MLBA
0.12287859	CZE_Hallstatt_Bylany
0.13108274	DEU_MA
0.13183519	CZE_Early_Slav
0.17643994	Berber_Tunisia_Chen
0.23509565	TUR_Ottoman
0.60633450	Dai
0.74575043	Yoruba

61.0	ITA_Rome_Imperial
16.6	ITA_Rome_Latini_IA
12.2	CZE_Early_Slav
5.2	TUR_Alalakh_MLBA
4.0	Berber_Tunisia_Chen
0.6	Dai
0.4	Yoruba

@Jovialis:

Would you take a WASP, a Jew, and an African-American, put them in one cohort, and call it "modern US"?

No, not now. But if they mix for 500 years and this becomes the main ancestral panmixed component in many regions of the USA, I would use it as a reference to estimate:
- Regional variation and deviation
- New immigration after the starting point of the panmixture.

I'm pretty sure that a slightly selected or even better new, additional samples would improve everything. But right now I think no other reference is better. And the reason for this is that in pre-Imperial times important ingredients were lacking and from Late Antiquity on, we see movements and new immigrants which created a new panel. So I would like to use a better reference for Romans. Which would you suggest? I'm ready to try

 

[brick]

TUR_Alalakh_MLBA is 50% Anatolian Neolithic, what should its use in models ever prove? It doesn't prove anything.

 

[Salento]

@Ygorcs

Dodecad K7b gives me R836 as my closest match, which is Italian, in your list. Which is very close to the two-way admixture I get with R1 and R850, when using just the Iron Age samples.

Here is what this calculator gives the Imperial samples I get for modern populations. I am not sure which "Greek" sample it is using, but I am sure if they had a better Italian set, it would put that instead:

Distance to: R836_Imperial_Era_Civitanova_Marche
3.75981383 Greek

Distance to: R835_Imperial_Era_Civitanova_Marche
2.66728326 Greek

Distance to: R113_Imperial_Era_Via_Paisiello_Necropolis
1.73568430 Greek

Distance to: R1549_Imperial_Era_Monterotondo
2.74479507 Greek

Distance to: R131_Imperial_Era_Via_Paisiello_Necropolis
2.39370006 S_Italian_Sicilian

Distance to: R49_Imperial_Era_Centocelle_Necropolis
4.15213198 S_Italian_Sicilian

Distance to: R47_Imperial_Era_Centocelle_Necropolis
3.12331555 S_Italian_Sicilian

As for the context of where they are from, they are not representative of Italy as a whole, considering they come from only two cemeteries outside of Rome.



It would be like someone using a cemetery in New York City, to determine the ethnic character of the United States.

.. highlighted in yellow are considered close to Apulian like.

... glad that none of my matches made the @Ygorcs ‘Foreign Tourists’ list

 

[Jovialis]

@Riverman, That looks wrong, even Ygorcs's list, using the same tool as you demonstrates that. Again, "Imperial Rome" is not a component, also it makes absolutely no sense to model samples from locations, and time periods that do not make sense for the target. Many of which have a lot of genetic overlap, as Brick pointed out.

There is absolutely no evidence of a "pan-mixture", and even if there was, the immigrants would get absorbed into the larger population. There wouldn't be any significant change to the ethnic character of the population. Even then, the study says there was deep demographic change in Late Antiquity. However, it is clear that all of the imperial samples that I get match first with R437 when using the Iron Age samples, and with modern Italian populations.

 

[Riverman]

TUR_Alalakh_MLBA is 50% Anatolian Neolithic, what should its use in models ever prove? It doesn't prove anything.

It proves that you need additional, more Eastern, Iranian shifted sources to account for a lack of Imperial_Roman as a more solid reference. Imperial Roman got it, but so do modern Italians, especially Central and Southern ones. Even if I would add GRC_Minoan_Lassithi, which in itself is already more Eastern shifted than EEF on the continent were, this won't go away. With the Imperial_Roman sample you have, exactly the way it is, a perfect stand in to account for most of the variation of the basic component in modern Central and Southern Italians. It is only slightly too much Levantine shifted, really only slightly so. But I have no better ancient reference and it works well.
How you would try to search for "Imperial Roman" ancestry in and outside of Italy? Because right not this seems to be the best proxy. I'm not saying identical, I'm not saying that at all, but a good proxy it is.

 

[Salento]

If I add my coordinates in the DodW9 ‘SOURCE’, ... I’m R437 closest match ... and Vice Versa!

Dod K7:

 

[Jovialis]

It proves that you need additional, more Eastern, Iranian shifted sources to account for a lack of Imperial_Roman as a more solid reference. Imperial Roman got it, but so do modern Italians, especially Central and Southern ones. Even if I would add GRC_Minoan_Lassithi, which in itself is already more Eastern shifted than EEF on the continent were, this won't go away. With the Imperial_Roman sample you have, exactly the way it is, a perfect stand in to account for most of the variation of the basic component in modern Central and Southern Italians. It is only slightly too much Levantine shifted, really only slightly so. But I have no better ancient reference and it works well.
How you would try to search for "Imperial Roman" ancestry in and outside of Italy? Because right not this seems to be the best proxy. I'm not saying identical, I'm not saying that at all, but a good proxy it is.

Again, there is no such thing as "Imperial Roman" ancestry, it makes no sense. It was different groups that lived in the same period of time. There is no proof that they mixed together, to change the ethnic character of the peninsula.

Also, what exactly composes Roman_Latini? Please don't tell me they lumped in all of the Iron Age samples together. There are only 4 "Latin" Tribe samples, R850, R851, R437, and R435. Half cluster with Southern Italians, the other half exist between Iberia and Northern Italy, along with Etruscan samples.

This is part of the reason why I have an issue with G25.

 

[Riverman]

@Riverman, That looks wrong, even Ygorcs's list, using the same tool as you demonstrates that. Again, "Imperial Rome" is not a component, also it makes absolutely no sense to model samples from locations, and time periods that do not make sense for the target. Many of which have a lot of genetic overlap, as Brick pointed out.

I'm sure better models can be made, but hardly without the Imperial Roman samples. Probably selecting them would improve it, but I have my doubts. I probably will try with a selection.

Also, I tried to find the best proxies and it seems to work out well enough, because I don't just compare Italians, but many European and West Asian populations with each other. If there is a big tilt where there shouldn't be one, I correct it.

There is absolutely no evidence of a "pan-mixture", and even if there was, the immigrants would get absorbed into the larger population.

Absolutely, that's my point all along. They were absorbed and caused the big "Mediterranean" component from the study, much, much bigger than anything between the Bronze Age and Imperial times, with a clear Eastern, Greek-like shift.

There wouldn't be any significant change to the ethnic character of the population.

That's debatable but if we start with that, it would be a complete different debate. Besides, most European people changed, continuity is the exception.

Even then, the study says there was deep demographic change in Late Antiquity.

Exactly. The trajectories completely changed. I'd say that from the point in time the empire was finally splitted, the centre moved to the East, the immigration from the East got drastically reduced and the immigration from the Western provinces and "Barbarians" drastically increased. Also the Northern Italian provinces got greater importance once again. This caused a backshift to a more Northern position. But most of the emigration from the East and Italy to the Northern and Western provinces, when this parts of the Empire influenced places like Britain, France, Germany etc., came in Imperial times and were, more likely than not, similar to the Imperial_Roman samples.

Again, there is no such thing as "Imperial Roman" ancestry, it makes no sense. It was different groups that lived in the same period of time. There is no proof that they mixed together.

Also, what exactly composes Roman_Latini? Please don't tell me they lumped in all of the Iron Age samples together. There are only 4 "Latin" Tribe samples, R850, R851, R437, and R435. Half cluster with Southern Italians, the other half exist between Iberia and Northern Italy, along with Etruscan samples.

This is part of the reason why I have an issue with G25.

But that's the kind of fruitful debate I want to have. Point me and others to what exactly might be wrong and could be improved. Actually, such samples can be selected and other sources put there instead. Its no big deal to correct such models. What would you take as a stand in for early Republican Romans?

 

[Jovialis]

^^personally, I think the calculator runs I posted back thread justify my position. The samples from the Imperial era, that I, Salento, and other Italians match with coincide with modern Italians, Ygorcs' list which uses G25, as well as consistently plot in the range of the two southern-Italian like Iron Age latins, R850, and R437.

We know that there was indeed deep demographic shifts in late antiquity. The immigrants, as has been argued by contemporary authors such as Brian Ward-Perkins, vanish from Italy, due to economic collapse, and devastation of war. To be replaced by native Italians from the hinterlands. My father's town for example was abandoned for 300 years, only to be repopulated in the middle ages.

I think while the Moots paper has some very interesting insights, we need more samples, especially from the south.

 

[Jovialis]

What would you take as a stand in for early Republican Romans?

I think R1 is representative of Italics, while R850 is more representative of a pre-Italic population, with R437 being intermediary. Of course there was admixture from both Etruscans, who are definitely more Northern Italian-like, but also Greek colonists, and Illyrians to consider.

 

[brick]

Why east sicily and campania
Score 13-14% levant ppnb in g25 runs by ygorcs
?
Maybe the levant signal is there and people
Don't want to see it and ignore it
I know people thinks mdlp k11 is not good
But please explain to me why other
Southern europeans like iberians and greeks
Score much lower % basal than south italians in mdlp k11?....

Levant_PPNB is almost 50% Anatolia_Barcin, so using Levant_PPNB in models could almost double what is the true value of Levant.

 

[brick]

It proves that you need additional, more Eastern, Iranian shifted sources to account for a lack of Imperial_Roman as a more solid reference. Imperial Roman got it, but so do modern Italians, especially Central and Southern ones. Even if I would add GRC_Minoan_Lassithi, which in itself is already more Eastern shifted than EEF on the continent were, this won't go away. With the Imperial_Roman sample you have, exactly the way it is, a perfect stand in to account for most of the variation of the basic component in modern Central and Southern Italians. It is only slightly too much Levantine shifted, really only slightly so. But I have no better ancient reference and it works well.
How you would try to search for "Imperial Roman" ancestry in and outside of Italy? Because right not this seems to be the best proxy. I'm not saying identical, I'm not saying that at all, but a good proxy it is.

The first traces of Iran_N were found in the Neolithic age in Italy and there is a complete lack of samples from the Iron and Bronze Ages in southern Italy in order to draw conclusions.

Considering that Davidski is constantly modifying and adjusting his averages and values of ancient and modern samples on G25 so that are in line with his ideas, considering that these instruments are very inaccurate, and that we are talking about minimum percentages, which become really significant only in the extreme south of Italy, it seems to me quite strange that many users are only interested in discussing this when it comes to Italians genetics.

 

[Leopoldo Leone]

The first traces of Iran_N were found in the Neolithic age in Italy and there is a complete lack of samples from the Iron and Bronze Ages in southern Italy in order to draw conclusions.

Considering that Davidski is constantly modifying and adjusting his averages and values of ancient and modern samples on G25 so that are in line with his ideas, considering that these instruments are very inaccurate, and that we are talking about minimum percentages, which become really significant only in the extreme south of Italy, it seems to me quite strange that many users are only interested in discussing this when it comes to Italians genetics.

I didn't know that G25 samples and averanges were curated by Davidsky, are you sure? If that is true, that is another strong reason why the results for Italians by that software ought not to be trusted when they are in contrast with the literature, given the explicit biases he manifested.

 

[Palermo Trapani]

The first traces of Iran_N were found in the Neolithic age in Italy and there is a complete lack of samples from the Iron and Bronze Ages in southern Italy in order to draw conclusions.

Considering that Davidski is constantly modifying and adjusting his averages and values of ancient and modern samples on G25 so that are in line with his ideas, considering that these instruments are very inaccurate, and that we are talking about minimum percentages, which become really significant only in the extreme south of Italy, it seems to me quite strange that many users are only interested in discussing this when it comes to Italians genetics.

Kudo's to you, good to see the American-Anglo alliance is still in tact. I appreciate your comment here.

 

[Riverman]

The first traces of Iran_N were found in the Neolithic age in Italy and there is a complete lack of samples from the Iron and Bronze Ages in southern Italy in order to draw conclusions.

Considering that Davidski is constantly modifying and adjusting his averages and values of ancient and modern samples on G25 so that are in line with his ideas, considering that these instruments are very inaccurate, and that we are talking about minimum percentages, which become really significant only in the extreme south of Italy, it seems to me quite strange that many users are only interested in discussing this when it comes to Italians genetics.

Iran_N was present earlier, but usually (wo. outliers) on a very low level. I agree that the lack of samples is a pity and am happy to embrace new ones. In the meantime, concerning accuracy: I looked up some of the Italian samples which being closest to me, not just by G25, but also by mytrueancestry and other tools available. Seems there is a pattern:



The good thing about these samples is that some of them seem to be, in my humble opinion, recent mixtures (!), so you can really check what's going on.

ITA_Rome_Late_Antiquity:RMPR106 looks like being 3/4 (Eastern? Gothic? w. slight Slavic?) Germanic and 1/4 Roman.
ITA_Rome_MA:RMPR1288 looks like 1/2 Germanic, 1/4 Slavic and 1/4 Roman.
ITA_Collegno_MA:CL63 looks like being 1/2 Germanic, 1/4 Celtic and 1/4 Roman (alternatively 3/4 Germanic w. slight Celtic and 1/4 Roman).
ITA_Rome_Renaissance:RMPR1219 looks like being 1/2 Germanic, 1/4 Slavic and 1/4 Roman.

To me that's quite interesting and its not like these results are completely off. I would never go for every single percentage in big runs, especially not with such diverse samples like Imperial Roman. But if you have the right references, like in this case, you get meaningful results. You can read up on the sample Collegno 63. CL63 is in the Northern (Germanic) category, but he is not above 90 percent cut off, only up to 70 - compare Table S14.1
https://www.biorxiv.org/content/biorxiv/suppl/2018/02/20/268250.DC1/268250-1.pdf

That's because one of his grandparents is supposedly Roman (provincial Pannonian? provincial Northern Italian?) as you can also check with the G25 tools.

 

[brick]

I didn't know that G25 samples and averanges were curated by Davidsky, are you sure? If that is true, that is another strong reason why the results for Italians by that software ought not to be trusted when they are in contrast with the literature, given the explicit biases he manifested.

Of course I am sure, it is a known fact that the G25 was invented by Davidski, and anything on G25 is managed and controlled by Davidski (aka Eurogenes aka Polako).

The results of the G25 regarding Italians are not completely in contrast with academic studies, we must be honest. In Italy the further south you go, the more genetics contributions are found from the eastern Mediterranean as showed also by academic studies. Except that Davidski seems to have a bias towards the Italians and therefore the results of the Italians seem to exaggerate the percentages of some components, and these components instead disappear in other non-Italian ethnic groups that should have them.

 

[Riverman]

Of course I am sure, it is a known fact that the G25 was invented by Davidski, and anything on G25 is managed and controlled by Davidski (aka Eurogenes aka Polako).

Doesn't mean the tools are bad or the references rigged. To say so would be quite an accusation. But even if true, what I would deny, the results are still solid as the simple run from above proves.

We can extend the samples I provided above with others from Pannonia:
HUN_MA_Szolad:SZ2: Fully Germanic
HUN_MA_Szolad:SZ22: 3/4 Germanic, 1/4 between Eastern Celts and Slavic (local population)
HUN_MA_Szolad:SZ23: 1/2 Germanic, 1/4 between Eastern Celts and Slavic (local population), 1/4 Imperial Roman citizen

Especially in cases of recent mixture, its absolutely clear as daylight and it works, even with such simple runs. And you can again compare with the study on the Szolad samples, they won't negate the results. The Roman component either scores as Imperial_Roman or some Italic sample + Levantine-Anatolian admixture. Always the same in these samples.

 

[brick]

Doesn't mean the tools are bad or the references rigged. To say so would be quite an accusation. But even if true, what I would deny, the results are still solid as the simple run from above proves.

No wonder you're acting like Davidski's defense attorney. I'm not saying the G25 is completely wrong in its results, I'm saying it can't be considered completely accurate. It's bad that anyone who thinks otherwise is banned from Anthrogenica.


I'm struggling to understand your point but I imagine it has some connection to what happened during Imperial Rome.
​

 

[Riverman]

I'm struggling to understand your point of view, but I imagine it has some connection to what happened during Imperial Rome.
​

I'm just defending the position I deem to be right, as long as I don't have the proof for it being wrong - in case of which I would change my opinion.

You said one cannnot work with the Imperial Roman sample, which is questionable, even if I agree that improvements and more samples would be - of course - just preferable and better. I also used the calculator in samples which are more straight up and you see all the results make perfect sense and align well with the results from the scientic studies which provided them. I mean up to the exact percentage and position on the PCA everything is right. And you get these "Imperial_Roman" profile not just in Italy and the city of Rome, you have it in Pannonia/Szolad, in Collegno/Northern Italy and many other places. It was spread in Imperial Roman times.

 

[Ygorcs]

Considering that Davidski is constantly modifying and adjusting his averages and values of ancient and modern samples on G25 so that are in line with his ideas, considering that these instruments are very inaccurate, and that we are talking about minimum percentages, which become really significant only in the extreme south of Italy, it seems to me quite strange that many users are only interested in discussing this when it comes to Italians genetics.

My models are not based on averages, but on the sum of all individual samples. Or are you saying he also deliberately changes the coordinates of individual samples just to fit his ideas?

We could easily test this. If some South Italian member of this or other forums got his/her G25 coordinates, we could easily run the same models and see if it deviates strongly from the South Italian samples in G25 spreadsheets.