|Forum||Europe Travel Guide||Facts & Trivia||Genetics||History||Linguistics|
|Eupedia Top > Eupedia Forum & Japan Forum|
Last edited by John Doe; 23-07-14 at 14:51.
Here are the Eurogenes K15 results for AJs and West-Sicilians:
Population Ashkenazi West_Sicilian Norwegian French Danish North_Sea 9.89 9.763333333 39.74 28.25 36.56333 Atlantic 10.7 18.32 23.47 26.05 27.75667 Baltic 6.74 4.696666667 13.26 8.22 11.58667 Eastern_Euro 5.17 3.27 11.48 6.32 10.71333 West_Med 14.82 17.75666667 6.36 15.53 5.99 West_Asian 13.05 10.64333333 2.24 4.66 3.34 East_Med 26.61 26.74666667 0.8 6.72 2.036667 Red_Sea 8.07 5.373333333 0.14 2.83 0.29 South_Asian 1.2 0.566666667 0.81 0.71 1.026667 Southeast_Asian 0.6 0.19 0.08 0.09 0.08 Siberian 0.58 0.083333333 0.5 0.09 0.093333 Amerindian 0.32 0.01 0.72 0.2 0.186667 Oceanian 0.24 0.52 0.32 0.15 0.073333 Northeast_African 1.63 1.326666667 0.06 0.11 0.186667 Sub-Saharan 0.39 0.723333333 0.04 0.07 0.07
The two components that the West-Sicilians most clearly have more of are the Atlantic and West_Med components. This seems plausibly to be in alignment with the theory of them having Norman admixture, since those two components are very West-European, and are carried in fairly high proportions among the French, Norwegian and Danish populations, from which the Normans are said to be descended. The components in order where the AJs are most clearly higher are: Red_Sea, West_Asian, Baltic, and Eastern_Euro, and they are not much higher in those components than the West_Sicilians. The North_Sea component is marginally higher in the AJs. The largest difference is clearly in the Atlantic component. In light of this, maybe I wouldn't rule out just yet the AJs having a bit of that Khazar admixture - albeit a very small proportion. Here's the source table: https://docs.google.com/spreadsheet/...Hc&usp=sharing
AJs are actually closer to Eastern Sicilians on the Eurogenes amateur crap.
Here are my K15 results:
Population North_Sea 8.44% Atlantic 19.52% Baltic 4.00% Eastern_Euro 1.99% West_Med 14.60% West_Asian 16.32% East_Med 26.20% Red_Sea 6.79% South_Asian - Southeast_Asian 0.62% Siberian - Amerindian - Oceanian 0.29% Northeast_African 1.24% Sub-Saharan -
I seem to get double the Atlantic component than an average AJ, but I suppose that just like Joey commented above, Gedmatch isn't very reliable. Is it possible that the similarities between AJs and west Sicilians come from a common Canaanite ancestry (the Phoenicians colonized west Sicily and like the Israelites they were Canaanites)?
Please feel free to correct me if I'm wrong.
Yeah pretty much that. Eastern Euro Ashkenazis have also a small amount of Turkic ancestry, that Sephardim lack.
A first-hand "census" on Jewish populations is given by Benjamin of Tudela's account of his travel around the Mediterranean in the second half of the 12th century. The figures probably relate to households rather than individuals. Around the Central/ Eastern Mediterranean, the largest Jewish communities he lists are:
- Alexandria 3,000
- Constantinople: 2,000 Rabbinites, 500 Karaites
- Thebes 2,000 (silk-weaving community, partly resettled by Normans into Sicily prior to Benjamin's journey)
- Palermo 1,500
- Salerno 600
- Naples, Otranto, Thessaloniki, Tyros: 500
Other relevant South Italian cities he visited included Capua (300), Taranto (300), Benevento (200), Melfi (200), Messina (200) and Trani (200).
For reference: Rhodes 400, Marseille, Corinth 300, Rome, Jerusalem 200, Lucca 40, Antioch 10, Genoa 2.
All those South Italian Jews must have gone somewhere after their expulsion in the 14th/ 15th century. And the communities were certainly ancient - we are talking about old Greek colonies here.
@joeyc: Yes, I didn't see the East-Sicilians in the table there. And yes, the East-Sicilians (ES) do look closer to the AJs than the West-Sicilians - quite similar actually. The largest differences between the AJ and ES in order are the Eastern_Euro, Atlantic, and West_Med components. Some might not find the table worthwhile, but I feel the percentages are worthy of posting anyway:
Population East_Sicilian Ashkenazi West_Sicilian North_Sea 9.75 9.89 9.763333333 Atlantic 13.27 10.7 18.32 Baltic 6.393333333 6.74 4.696666667 Eastern_Euro 2.26 5.17 3.27 West_Med 17.35666667 14.82 17.75666667 West_Asian 14.43 13.05 10.64333333 East_Med 26.94333333 26.61 26.74666667 Red_Sea 6.763333333 8.07 5.373333333 South_Asian 0.52 1.2 0.566666667 Southeast_Asian 0.03 0.6 0.19 Siberian 0.003333333 0.58 0.083333333 Amerindian 0.063333333 0.32 0.01 Oceanian 0.25 0.24 0.52 Northeast_African 1.596666667 1.63 1.326666667 Sub-Saharan 0.36 0.39 0.723333333
But even a AJ/ Sephardim mix wouldn't make you that "Atlantic", Sephardim also have just 14% of it. Actually, there aren't many populations that are high on "Atlantic" but at the same time low on "North Sea". The best here is French Basques (45% Atlantic vs. 17% North Sea) - in terms of a relevant, ancient Jewish community that should mean Narbonne.
The "best fit" population, as you probably already have realised, is West Sicilians. But West Sicilians, and even more so French Basques lack the West Asian component and are too Western Mediterranean. So I looked for a population that is high on "West Asian" but at the same time low on "Eastern European" and "Siberian". Best fit here is Georgian Jews. From the three together one can already pretty well emulate your ancestry structure, but they miss an East African & Red Sea component. Yemenite Jews fill the gap perfectly. That's not meaning you have actually Yemenite Jewish ancestry, they should rather compensate for the specifically Jewish element that is missing within French Basques. So here is my result:
Population Georgian Jewish West Sicilian Yemenite Jewish French Basque Mix John Doe 35.6% 26.1% 9.1% 29.2% 100% North_Sea 2.27 9.76 0.27 16.85 8.30 8.44 Atlantic 2.67 18.32 1.05 45.40 19.08 19.52 Baltic 1.63 4.70 0.16 4.79 3.22 4.00 Eastern_Euro 2.91 3.27 0.11 2.82 2.72 1.99 West_Med 6.95 17.76 4.40 25.04 14.82 14.60 West_Asian 35.22 10.64 5.41 0.83 16.05 16.32 East_Med 36.66 26.75 54.11 2.74 25.76 26.20 Red_Sea 8.06 5.37 27.46 0.73 6.99 6.79 South_Asian 2.77 0.57 0.48 0.13 1.22 0.00 Southeast_Asian 0.17 0.19 0.06 0.28 0.20 0.62 Siberian 0.10 0.08 0.10 0.04 0.08 0.00 Amerindian 0.16 0.01 0.08 0.03 0.08 0.00 Oceanian 0.11 0.52 0.18 0.10 0.22 0.29 Northeast_African 0.29 1.33 6.11 0.17 1.06 1.24 Sub-Saharan 0.04 0.72 0.01 0.04 0.21 0.00
Except for the South Asian component that comes with the Georgian Jews, it fits quite nicely.
Otherwise, you appear to be a living proof of Jewish settlement in Eastern Europe (Galizia) during Khazar times.
All Jews were completelly expelled from Southern Italy in the XVI century, including all the half breeds and the converts.
Their descendants would become the Italkim of central and northern Italy. All the Sephardim and Ashkenazim of Italy lived in the North (Rome, Trieste, Livorno, Venice....)
AJs are much closer to Eastern Sicilians (Siracusa) than to Western ones (Trapani). Which is funny considering that Siracusa was never occupied by Phoenicians and the Arab influences were quite insignificant.
As for the French Basque, well look at my K15 results:
Admix Results (sorted):
# Population Percent
1 East_Med 26.2
2 Atlantic 19.52
3 West_Asian 16.32
4 West_Med 14.6
5 North_Sea 8.44
6 Red_Sea 6.79
7 Baltic 4
8 Eastern_Euro 1.99
9 Northeast_African 1.24
10 Southeast_Asian 0.62
11 Oceanian 0.29
Single Population Sharing:
# Population (source) Distance
1 Ashkenazi 9.1
2 Sephardic_Jewish 9.48
3 Tuscan 10.37
4 Greek 10.57
5 North_Italian 16.43
6 Bulgarian 16.92
7 Cyprian 17.15
8 Turkish 18.23
9 Romanian 18.75
10 Lebanese_Muslim 19.17
11 Syrian 20.2
12 Serbian 21.68
13 Spanish_Andalucia 22.16
14 Spanish_Extremadura 22.31
15 Spanish_Murcia 22.54
16 Jordanian 23.15
17 Portuguese 23.16
18 Samaritan 23.47
19 Spanish_Valencia 23.63
20 Spanish_Castilla_Y_Leon 23.85
Mixed Mode Population Sharing:
# Primary Population (source) Secondary Population (source) Distance
1 64.8% Lebanese_Muslim + 35.2% French_Basque @ 2.51
2 57.8% Lebanese_Muslim + 42.2% Spanish_Aragon @ 3.25
3 55.9% Lebanese_Muslim + 44.1% Spanish_Castilla_La_Mancha @ 3.28
4 51.9% Assyrian + 48.1% Spanish_Aragon @ 3.35
5 55.3% Lebanese_Muslim + 44.7% Spanish_Valencia @ 3.43
6 53.7% Lebanese_Muslim + 46.3% Spanish_Andalucia @ 3.48
7 59.4% Lebanese_Muslim + 40.6% Southwest_French @ 3.52
8 52.3% Spanish_Andalucia + 47.7% Assyrian @ 3.54
9 50% Spanish_Castilla_La_Mancha + 50% Assyrian @ 3.58
10 50.6% Spanish_Valencia + 49.4% Assyrian @ 3.72
11 59.3% Assyrian + 40.7% French_Basque @ 3.81
12 58.6% Lebanese_Muslim + 41.4% Spanish_Cantabria @ 4
13 51.9% Spanish_Murcia + 48.1% Assyrian @ 4.38
14 54.2% Lebanese_Muslim + 45.8% Spanish_Murcia @ 4.4
15 55.7% Lebanese_Muslim + 44.3% Spanish_Castilla_Y_Leon @ 4.43
16 53.5% Assyrian + 46.5% Southwest_French @ 4.44
17 56% Lebanese_Muslim + 44% Spanish_Cataluna @ 4.47
18 63.8% Syrian + 36.2% French_Basque @ 4.5
19 52.7% Assyrian + 47.3% Spanish_Cantabria @ 4.69
20 50.4% Spanish_Castilla_Y_Leon + 49.6% Assyrian @ 4.71
As for the Khazar admixture, maybe, but that still wouldn't explain so much Atlantic admixture. Also I highly doubt I have Yemenite or even Georgian Jewish ancestry, in the case of the former, I'd expect much more red sea and a much closer proximity to Palestinians, Jordanians, Bedouins and Saudis, with Yemenite Jews being Judaised Arabs and all.
Your turkic ancestry is most likely hided in the Baltic and Eastern Euro clusters AFAIK.
@Joh Doe: In the case of French Basque/ Atlantic admix I of course forgot Bordeaux and Toulouse as relevant and ancient Jewish communities. Both were within the Visigothic realm, which appear to have crafted a specific political alliance with the Jews. They didn't issue any discriminations, as Ostrogtothic Theodoric, and Jews are reported to have been among the most ardent supporters of Visigothic rule. Moreover, the region was part of the Norman/ English realm, and thus out of reach of earlier Medieval French discrimination and prosecution during the First Crusade. Especially Jews from Tolouse, however, became implicated in French prosecution of the heresy of the Albigenses in the 13th century - Alsace / Switzerland, and in the mid-14h century from there to Eastern Europe looks like a plausible migration path. Otherwise, Southern Italy with its high religious tolerance under 13th century Hohenstaufen rule would probably be the place to go if you wanted to escape the inquisition yet continue trade in Mediterranean commodities.
Eurogenes K15 West Asian admix centres in on the Greater Caucasus. After 200 years of Russian control, all North Caucasian populations have picked up some 6-10% East European ancestry, which makes them a rather poor fit for your case. So I took the Georgian Jews as a proxy of a lowly-russified Caucasian ancestry. Such ancestry appears to be stronger present with you then with the average AJ - whether due to direct Georgian-Jewish ancestry, or some Caucasian-Khazarian-Jewish ancestry that was picked up in Galizia, is something my amateurish approach cannot tell.
Noting from your Eurogenes admix results that Lebanese Muslims appear to capture the original Jewish (East Med/ Red Sea/ East African) ancestry component quite well, I have run another simulation where Yemenite Jews are replaced by Lebanese Muslims. The caveat here is that Lebanese Muslims have a bit of South Asian and Sub-Saharan African admix. Both tend to somewhat reduce the component's weight, but are probably not a fair representation of how the Levante's genetic mix looked 2,000 years ago. So I constructed an artificial Lebanese Muslim structure where these components are proportionally distributed onto all other components. Since that should capture the Jewish component, I also replaced Georgian Jews by Georgians. Moreover, as your father looks Anglo-German, I took in a fourth component. I have experimented with various options (Dutch, English, West German etc.), Northern German works best for its elevated Baltic admix.
That still left one problem - that 0.6% SE Asian admix in your genes, and with the average AJ, which is neither European nor Levantine. Thus, I looked for a population that maximises SE Asian ancestry with at little as possible Siberian and South Asian admix. Best choice here are the Dai (SW China), though in fact we are probably talking about some Himalayan component transferred via Afghanistan / Central Asia. Its highest frequencies in Europe (1.5-3%) are found among some North Caucasian populations, Ukraine is 0.5% (though Belgorod/Lviv only 0.2%). As such, the Georgian-Dai mix may be taken as a proxy of how the North Caucasian (Khazar) gene pool looked prior to Russification.
Here is the result (f²=0.46):
Population Lebanese Muslim 2 French Basque Georgian North German West Sicilian Dai Mix John Doe 58.0% 32.1% 4.3% 3.7% 1.6% 0.4% 100% North_Sea 2.50 16.85 4.68 33.08 9.76 0.17 8.44 8.44 Atlantic 5.79 45.40 3.90 27.46 18.32 0.17 19.39 19.52 Baltic 2.15 4.79 3.69 13.41 4.70 0.29 3.52 4.00 Eastern_Euro 1.26 2.82 3.27 9.95 3.27 0.29 2.20 1.99 West_Med 10.06 25.04 4.64 6.11 17.76 0.12 14.57 14.60 West_Asian 22.94 0.83 53.85 5.24 10.64 0.20 16.24 16.32 East_Med 41.06 2.74 19.63 2.50 26.75 0.15 26.04 26.20 Red_Sea 11.04 0.73 3.24 0.70 5.37 0.07 6.88 6.79 South_Asian 0.13 2.12 0.83 0.57 2.70 0.18 0.00 Southeast_Asian 0.12 0.28 0.16 0.02 0.19 91.41 0.52 0.62 Siberian 0.09 0.04 0.23 0.08 0.08 2.02 0.09 0.00 Amerindian 0.04 0.03 0.34 0.22 0.01 0.57 0.06 0.00 Oceanian 0.56 0.10 0.08 0.14 0.52 1.47 0.38 0.29 Northeast_African 2.40 0.17 0.13 0.19 1.33 0.17 1.48 1.24 Sub-Saharan 0.04 0.05 0.08 0.72 0.20 0.03 0.00
Except for the Baltic admix, which is still a bit too low, this looks quite good. The under-representation of the Baltic admix is related to its low share within Georgians. If I had taken North Caucasian populations (e.g. Adygei or Balkars) instead, the Baltic admix would have been fine (but in that case I would have gotten too much Eastern European admix).
I also checked which of the above components can be removed without significantly worsening the mix:
- West Sicilian: Only slight worsening (f²=0.48). Of the 1.6% West Sicilian, 1% goes to Lebanese Muslims, 0.6% to French Basques. The Atlantic component actually gets better, but West Med is now slightly under- and Red Sea bit more overestimated.
- Dai: Somehow worse (f²=0.56), as the SE Asian admix cannot anymore be approximated. The mix gets more Sicilian (2.3%), and less Basque/ Lebanese.
- North German: Substantially worse fit (f²=1.29). The simulation is now pushing up Sicily to 12%, Georgians up to 6% and Lebanese Muslims down to 51%,which improves a bit the East Med/ Red Sea/ East African components. However the mix becomes far too West Mediterranean and underestimates the North Sea and Baltic components by 0.5-0.7%.
- Georgian: Far worse. In order to somehow fit your West Asian admixture, the model tries to maximise the Lebanese Muslim component to such an extent that the Sicilian component turns negative. The best valid result has an f² of 2.84, it maximises Lebanese Muslims at 63%, and reduces the Sicilian component to zero. Still, the West Asian component is clearly underrepresented, Red Sea gets too high. Otherwise, it doesn't work too badly. By pushing up the Northern German component to 5.6%, the European components (Baltic, North Sea, Atlantic, Western Med) are reasonably well approximated, though North Sea and Eastern Europe are now also getting slightly too high.
Bottom line: The Sicilian component can be eliminated without significantly worsening the mix, the others are required. Essentially, your ancestry can be modelled as 59% Levantine, 32.5% Aquitaine (French Basque), 4.5% Khazarian (Georgian plus additional SE Asian & Baltic admix) and 4% North German. Alternatively to North German, Southern Dutch isn't a bad fit either(f²=0.56), but misses a bit on the Baltic component; Northern Dutch are a bit too North Sea (f²=0.54).
That North German ancestry may of course have been picked up somewhere in Eastern Europe from a German colonist - say 5 generations ago. Otherwise, if some of your ancestors lived as Jews for a longer time in Medieval Germany, one could, aside from Cologne, consider the following places:
A few things.1) Only eastern AJs have a significant amount of supposed Khazar ancestry. Georgians are not a good proxy for the Khazars.2) The Sephardi cluster is made up of Bulgarian and Turkish Sephardim now AFAIK. The original Behar paper from 2009 had the Belmonte Jews from Portugal, who better represent the original Sephardim.
If my rudimentary Italian doesn't deceive me, it says that some 6% of Sicily's population in 1492 were Jews. Out of which total? Surely more than half a million, possibly one million (Sicily's population today is around 5 million, in spite of heavy emigration to the US and elsewhere). That means that at least 50,000, possibly 100,000 or more Jews had to emigrate from Sicily and Southern Italy. I furthermore understand that the stop-over in Central/ Northern Italy was only short, and after 1580 many (most?) of the Sicilian / Southern Italian Jews moved on to Poland (probably meant to include Galicia) and Lithuania.Si stima che nel 1492 gli ebrei componessero oltre il 6% della popolazione della Sicilia. Molti ebrei siciliani inizialmente andarono in Calabria, che già aveva una comunità ebraica sin dal IV secolo. Nel 1524 gli ebrei furono espulsi dalla Calabria e nel 1540 da tutto il Regno di Napoli, poiché queste regioni caddero sotto il dominio degli spagnoli e furono oggetto dell'editto di espulsione dell'Inquisizione spagnola.
Ci fu uno spostamento graduale degli ebrei durante tutto il XVI secolo dal sud d'Italia verso il nord, con il peggioramento delle condizioni per gli ebrei a Roma dopo 1556 e a Venezia negli anni 1580. Molti ebrei da Venezia e aree circostanti emigrarono verso la Polonia e la Lituania in questo period.
That would explain why AJ plot so close to Sicilians / South Italians. I'd guess if you "mix" West Sicilians with South Italians (especially Calabrians), the result should genetically resemble East Sicilians.
Almost all European/Turkish/North African Jews are the descendants of the expelled Iberian Jews actually. There was even a study regarding to Jewish Y-Dna supporting that.
The Khazar contribution is greatly overrated & overestimated.
Moreover, I find most Gedmatch runs wholly unconvincing given the nature of the components (derived from actual components such as WHG, EEF, ANE, ASI, ANI, etc).
Which is why I'm kind of amused when someone claims that "Eastern Ashkenazi Jews" have more "Khazar" or "Slavic" ancestry than "Western AJs". In fact, one of the Behar et al. 2013 paper's conclusions was that "relatively little observable genetic difference exists between representatives of eastern and western Ashkenazi Jewish populations, suggesting that genetically, the Ashkenazi Jewish population approximates a single large community (Guha and others, 2012)".
Regarding the Basque, I think you might be onto something here (no indication of gene-flow though, so the ~65% Lebanese figure is wrong, same story with the ~50% Iberian fits):
The above is an IBD map taken from Vadim Verenich's blog. As you can see, there's a non-negligible amount of IBD sharing with Basques... Along with Sardinians.
IMO this might have something to do with the fact that EEF-like ancestry was much more common around the first couple of centuries CE and managed to remain until the Ashkenazi bottleneck took place.
There's also high IBD sharing with the Greeks, so I'm more & more enclined to think that the Aegean played a big part in the formation of Western (Ashkenazi-Sephardi) Jewry.
In fact, I'm ready to say that some of the Aegean admixture might even go back to the Middle Bronze Age, when contacts between the Minoans (and Mycenaeans later on) and the Levant were thriving. This definitely isn't good news since this means that we'll be grasping at straws and splitting hairs when pre-exilic genome-wide results come in, as it strengthens the case for a Cypriot-like Levantine population prior to the emergence & spread of the arabs (which had already started a few centuries prior to the appearance Islam, mind you).
There's also a noticeable degree of IBD sharing with Eastern Ukrainians, in this case I think it has to do with prolonged cohabitation and the direction of gene-flow would be mostly Jewish to non-Jew given the region's history (replete with examples corroborating such a model, think of the Jewish cossacks for instance) and the paucity (not to say absence) of WHG in Jews.
Also of interest is the level of IBD sharing with Iranians, possibly a remnant of the Babylonian exile (would explain why I have Iranian and Uzbek Jewish relatives on my RF).
The Khazar theory takes yet another blow here, given the low to non-existent amounts of IBD sharing between Turkic speakers and Jews (save Uyghurs and Anatolian Turks, which are special species in their own right, so to speak).
Finally, I can't help but notice the fact that the level of IBD sharing with Italians is pretty low (lower than the amount of IBD sharing with Basques or Greeks, and on par with the level of IBD sharing with Turks & Iranians), which is surprising given the high similarity to Eastern Sicilians and Maltese (though in a sense you could say that AJs are even more similar to Aegean islanders and Cretans, but that's another story).
Oh, and Jewish uniparental lineages seem to be Near Eastern for the most, and yes that includes mtDNA as well.
So what you're saying is that the large IBD sharing between AJs and Greeks could go as far back as the middle bronze age due to contact between the Minoans and Mycenaeans on one hand and Levantines on the other, and later on through the Hellenistic and Roman periods?
The relation to eastern Ukraine could be also from that fact that many Jewish women were raped by Cossacks.
P.S Talking about Anatolian Turks, don't forget that Anatolian Turks aren't exactly Turkmen, I reckon they largely descend from Indigenous Anatolians who were Turkefied and Islamised by the actual Turks, just like modern Hungarians represent a largely indigenous population conquered and assimilated by a Magyar elite.
P.P.S Yeah I'm aware of the fact that Gedmatch is hardly as reliable as Behar and Lazaridis.
P.P.P.S Where did you get that map from?