I'm ready to bet that the only reason for the gap between Tunisian/Moroccan/Libyan Jews and Ashkenazim is entirely due to North African admixture.
Indeed, if you include NA populations (such as Mozabites) there's a clear cline of NA Jews towards these populations, suggesting admixture which makes sense given that Tunisian & Libyan Jews were amongst the first Jews to settle this area, with the Carthaginians thus giving them plenty of time to mix with neighbouring Berbers (there was a Judaic trend for quite some time and many Berber tribes eventually converted to Judaism).
This is easily picked up by most calculators and there is a non-negligible amount of IBD sharing between NA Jews and Mozabites for instance.
Moroccan & Algerian Jews seem to have absorbed much less admixture and might have more actual Sephardic ancestry per se.
Below is a representation of Lazarides' admixture results, copied from
http://www.eupedia.com/forum/thread...-Aryan-split-of-IE/page11?p=434227#post434227 (poster Alan).
I am not sure where it comes from. It is not part of Lazarides' original study, the colouring differs from Lazarides' admixture graphs, and the legend uses terminology (yDNA hgs) not used by Lazarides. Maybe Alan wants to give more detail on his source. Nevertheless, visual comparison suggests that the admixtures are at least very similar to those in Lazarides' original plots, so for the time being I take it as a fair representation of Lazarides' results.
And, in fact, the main difference between AJ and those from the South Mediterranean appears to be that the latter have picked up quite some Mozabite genes. AJ also have a bit of Mozabite genes, but far less. Instead, they have picked up quite some North European genes (mid-blue, peak in Fenno-Scandians and Balts), plus a tiny dose of dark blue North Eurasian genes (peak in Nganasan). The latter suggests some, very limited additional admixture in Eastern Europe. The Fenno-Scandian component is extremely close to the one found in Sicilians and Maltese, and may relate to various admixtures in the region, including (but most likely not restricted to) the Norman conquest of Sicily in the 10th century AD.
Turkish Jews basically haven't changed much since their ancestors left the Iberian peninsula, and they overlap with French & Ashkenazi Jews accordingly so.
If I had to guess, I'd say that the diaspora's forefathers plotted around Turkish Jews, who happen to overlap with Ashkenazim.
If you put all Western Jewish populations on a PCA plot, there's a clear cline of all Jewish groups towards Lebanese-leaning Cypriots... And if you ask me, I think that's where pre-exilic Judeans plotted.
I think you are overlooking the following:
http://en.wikipedia.org/wiki/Expulsion_of_the_Jews_from_Sicily
At the time of expulsion from Sicily, the Jewish community in Sicily dated back to early Roman times, and they were relatively untroubled on the island until the acceptance of the Crown of Aragon in Sicily in 1412. A great number of Jews had reached Sicily after Pompey's 63 BC sacking of Jerusalem, and additionally by Roman Proconsul Crassus, who is traditionally said to have sold more than 30,000 Jewish slaves on the island.
After the enslavement under Roman rule, Jews in Sicily eventually assimilated into society, working in professions such as philosophy, medicine, artisanal pursuits, and farming.
The exact number of Jews in Sicily at the time of expulsion is not certain, However, some have put the number of Jewish refugees at 36,000.[1] Also, in 1492, it is known the Jewish populations of Palermo, Messina, and several other cities were considerable, and that there were Giudeccas, or Jewish settlements, in over 50 places in Sicily, ranging in anywhere population from 350 to 5,000. At their height, Jewish Sicilians probably constituted from five to eight percent of the island's population.[2]
In 1492, as part of an attempt to maintain Catholic orthodoxy and purify their kingdom of Moorish influence, Ferdinand and Isabella ordered the forced expulsion or conversion of all Jews on pain of death. The date of the expulsion was extended from 18 September 1492 to 12 January 1493, in order to allow the extortion of opportunist tax levies. Many Sicilian Jews fled to neighboring Calabria where the Spanish Inquisition caught up with them again fifty years later. Not all of the Sicilian Jews departed. A large number of Sicily's Jews converted to Catholicism and remained on the island.
Actually, the Jewish population share might have been much larger. There is a high medieval petition by Jews from a Sicilian city (Catania? - I really should bookmark everything I read!) to adjust their tax due to their population share of 11%. In Southern Italy, especially Calabria, there have been many exclusively Jewish agricultural communities, and one paper puts the share of Jews among medieval Calabrians as high as 50%. Note in this respect that we are not talking "some Mediterranean island" here. Italian and German Wikipedia put Palermo's population under Arab rule (late 9th/ early 10th century) at 300,000. Other sources are more conservative, but there seems to be consensus that Arab Palermo had more than 100,000 inhabitants and equalled Byzantium in size. In spite of obvious population decline, mid-15th century Palermo is still believed to be among the top 10 European cities with a population close to 50,000. Naples had gained similar size by then, and Syracuse may not have been much smaller.
As is indicated by Paulus' missioning, Judaism enjoyed quite a popularity in the hellenisized population of the Roman empire. As such, it wouldn't be surprising if many Southern Italians had voluntarily adopted Judaism already by the time of Paulus' missioning. And, under Arab rule, it was probably better to be a Jew than to continue believing in ancient Greek gods. Or, put differently: If "Catholic" means Spanish/ Roman, "Orthodox" means Byzantine, and "Muslim" means Arab, what do you take for "None of the above" (e.g. former Greek colonists, or post-Vandal ex-Arianist, or "I want my good old Staufen emperors back, not these Spanish papist a..h..s")? Essentially, what I am trying to say is that there were probably hardly barriers to genetic exchange between Sicilian / South Italian Jews and non-Jews between antiquity and the end of Hohenstaufen rule around 1,300 AD, and Lazarides' results support this hypothesis.
http://en.wikipedia.org/wiki/History_of_the_Jews_in_Sicily
The exiles found protection under Ferdinand I of Naples in Apulia, Calabria and Naples. On the death of Ferdinand in 1494, Charles VIII of France invaded Naples. At that time a serious disease, known as "French fly," broke out in that region, and the responsibility for the outbreak was fixed upon the Jews, who were accordingly driven out of the Kingdom of Naples. They then sought refuge in Turkish territory, and settled chiefly in Constantinople, Damascus, Salonica, and Cairo. To remain in Sicily, a significant number of Sicily's Jewish population converted to Catholicism. Many of these converts remained Crypto-Jews, known as neofiti.
I think that explains why AJ/ Sicilians are quite close to Turkiish Jews, and especially the well-visible fenno-scandian and Mozabite elements in the latter's gene pool, both of which are almost completely absent in Mesopotamian / Caucasian Jews. Cypriots are quite different from Turkish Jews: Much more Caucaso-Gedrosian, no Maozabites, and just a little fenno-scandian element (which might, among others, relate to survivors from the Crusaders' states settling there).
Mizrahim seem to have absorbed a lot of Mesopotamian DNA, this is both apparent in their autosomal DNA & uniparental lineages. This is why they plot with Assyrians and Syriac Orthodox christians... Here again, we're dealing with communities which have been in Mesopotamia-Iran (and the Caucasus later on) for a very long time (at least since the Babylonian exile).
I do have Mizrahi relatives so they obviously have retained some of the original Israelite admixture... Still, I think the Mesopotamian contribution to Mizrahi Jews is greatly underestimated.
Some Georgian Jewish communities (e.g. the one in Kutaisi) claim direct descent from the first diaspora, and it is said that some Jews released from the Babylon exile went to Georgia instead of returning to Jerusalem. Considering archaeological evidence of copper trade from the Caucasus into today's Israel since almost 10,000 years, I think there is quite a likelihood of Mesopotamian and Caucasian Jewish communities having been established more or less simultaneously. There has obviously been strong interaction between both areas, not only during Khazar times, but also as Persians have over several periods in antiquity, and in the late Medieval, controlled part or all of the area south of the Greater Caucasus. Moreover, as I have reported in a previous post, a substantial number of Georgian Jews were deported into Iran in the 17th century. So, speculation on Caucasian-Mesopotamian genetic interchange, including the effects on the area's Jewish community, may turn into an "hen and egg" question. Lazarides' admixture analysis clusters them both as Caucaso-Gedrosian.
Iranians appear to incorporate some South Asian (light green) element, traces of which may be also spotted among Georgian and Iranian, but not Iraqian Jews. That would rather speak for gene flow from the Iranian plateau than from Mesopotamia into Mirzahi Jews. In any case, what distinguishes all Jewish populations most strongly from their host populations is elevated "Bedouin 2" genes. If I recall correctly, that would be Bedouins from the Sinai. Apparently, (proto-) Jews spent more than just 40 years walking through the desert.....
I'd be very cautious with any model implying high amounts of European ancestry, unless you can explain the insane paucity of WHG.
Indeed, Lazarides' results point at a pretty straight migration of AJ from Sicily into Eastern Europe (though it still would have been nice to have had some Rhineland data for comparison). Strangely, while the biographies of various noted South Italian Jewish scholars demonstrate emigration towards Northern Italy and the Eastern Mediterranean, there is hardly any historical evidence for emigration into Eastern Europe. Might we have had two distinct patterns - the South Italian Jewish urban elite settling in other larger cities, and the Jewish artisan/ farmers rather going into the (newly colonised) Eastern European countryside? That would still leave unanswered why AJ spoke Yiddish, and not some mix of Hebrew and South Italian.