May 4, 2014

Sicilian haploid genetics in the Mediterranean context

A new study takes a look at Sicilian haploid genetics in its wider geographical context.

Stephania Samo et al., An Ancient Mediterranean Melting Pot: Investigating the Uniparental Genetic Structure and Population History of Sicily and Southern Italy. PLoS ONE 2014. Open accessLINK [doi:10.1371/journal.pone.0096074]

Abstract

Due to their strategic geographic location between three different continents, Sicily and Southern Italy have long represented a major Mediterranean crossroad where different peoples and cultures came together over time. However, its multi-layered history of migration pathways and cultural exchanges, has made the reconstruction of its genetic history and population structure extremely controversial and widely debated. To address this debate, we surveyed the genetic variability of 326 accurately selected individuals from 8 different provinces of Sicily and Southern Italy, through a comprehensive evaluation of both Y-chromosome and mtDNA genomes. The main goal was to investigate the structuring of maternal and paternal genetic pools within Sicily and Southern Italy, and to examine their degrees of interaction with other Mediterranean populations. Our findings show high levels of within-population variability, coupled with the lack of significant genetic sub-structures both within Sicily, as well as between Sicily and Southern Italy. When Sicilian and Southern Italian populations were contextualized within the Euro-Mediterranean genetic space, we observed different historical dynamics for maternal and paternal inheritances. Y-chromosome results highlight a significant genetic differentiation between the North-Western and South-Eastern part of the Mediterranean, the Italian Peninsula occupying an intermediate position therein. In particular, Sicily and Southern Italy reveal a shared paternal genetic background with the Balkan Peninsula and the time estimates of main Y-chromosome lineages signal paternal genetic traces of Neolithic and post-Neolithic migration events. On the contrary, despite showing some correspondence with its paternal counterpart, mtDNA reveals a substantially homogeneous genetic landscape, which may reflect older population events or different demographic dynamics between males and females. Overall, both uniparental genetic structures and TMRCA estimates confirm the role of Sicily and Southern Italy as an ancient Mediterranean melting pot for genes and cultures.

No particular haplogroup is dominant in the island in the Y-DNA side and, although H has some clear prevalence among mtDNA haplogroups, it is actually well under the normal European levels for this common haplogroup.

Table 1. Age estimates (in YBP) of STR and HVS variation for the most frequent haplogroups in Sicily and Southern Italy.

Y-DNA

We can see how the following patrilineages are more common: J2a (16%), G2a (12%) and E1b1b1a1b1a (10%) and R1b1a2a1a2 (9%). R1a1a (5%), J1 (5%) R1b1a2a1a1 (4%) and J2b (4%) are less common instead.


Fig, S2(a) - Principal Component Analysis (PCA) based on haplogroup frequencies for Y-chromosome (a) (...). Population codes as in Table S1. Colour codes for geographic affiliations as in the legends at the bottom-left of each plot. Legend abbreviations: NAFR: North-Africa, LEV: Levant, BALK: Balkans, SSI: Sicily and South-Italy, NCI: North-Central Italy, IBE: Iberian Peninsula, GER: Germany.

There is an interesting tendency in Agrigento (AG) towards Lebanon (which in this graph includes all the LEV category), while other areas of Sicily and Southern Italy (Lecce, Cosenza, Enna) tend instead towards the Aegean (Pho, Smy). These tendencies could be interpreted (at least partly) in terms of historical colonization events by Phoenicians and Greeks. Catania instead tends towards Central-North Italy, maybe reflecting its important role under Roman rule and a historical colonization in the times of Augustus.

The Southern Italian towns of Matera (Basilicata) and Campobasso (Molise) also show a tendency towards the Northern Balcans (represented by Serbia here). 

The authors confirm previous impressions of a West-East Y-DNA duality in the Mediterranean that divides Italy:
When comparing SSI with Mediterranean reference populations, Y-chromosome results (Figure 1 and Figure S2) revealed a clear-cut genetic differentiation between the North-Western vs. the Central- and South-Eastern Mediterranean genetic pools (as confirmed by both sPCA G-test and AMOVA FCT statistically significant tests). These results are consistent with our previous study about Italy [12], in which we detected a discontinuous paternal genetic structure, clearly separating the South-Eastern and the North-Western parts of the Italian Peninsula. Here this pattern appears extended to the whole Mediterranean Basin, particularly suggesting a shared genetic background between South-Eastern Italy and the South-Eastern Mediterranean cluster from one side, and between North-Western Italy and the Western Europe from the other side (Figure 2).


Mitochondrial DNA

The main matrilinages of Sicily are H (28%) T (13%), J (10%) and HV(xH) (5%). U5 is also well under the usual European frequencies with just 3% of prevalence. 

While the AMOVA statistical significance tests say that PC2 in the following graph is not really significant. However PC1 is still relevant, I understand.

Fig S1(b) - Principal Component Analysis (PCA) based on haplogroup frequencies for (...) mtDNA (b). Population codes as in Table S1. Colour codes for geographic affiliations as in the legends at the bottom-left of each plot. Legend abbreviations: NAFR: North-Africa, LEV: Levant, BALK: Balkans, SSI: Sicily and South-Italy, NCI: North-Central Italy, IBE: Iberian Peninsula, GER: Germany.

If anything there is some discrepancy between Y-DNA tendencies and those of mtDNA. For example the "Phoenician" Agrigento in the Y-DNA graph, looks "Iberian" or "Tuscan" in the mtDNA one. 

The authors believe that mtDNA lineages could be older than Y-DNA ones in many cases:

Y-chromosome results however contrast with the lack of statistical support to the sPCA global structure observed for mtDNA diversity, excepted for a similar NW-SE genetic pattern identified by sPC1 (Figure 3). The common South-East to North-West pattern in the distribution of genetic variation across the European and Mediterranean domain, could be interpreted as reflecting the same SE to NW genetic cline extensively reported in literature for the whole of Europe [71][74]. However, the general lack of statistical support to the global structure observed for mtDNA markers suggests a higher homogeneity for maternal than paternal genetic pools in the Mediterranean genetic landscape. These results could be ascribed to older population events and/or different demographic and historical dynamics for females than males. (...) In fact, whereas the different continental and within continental contributions to the current SSI genetic pool appeared to be more equally distributed on the maternal side (despite a noteworthy contribution of Levantine females), the paternal counterpart appeared to be clearly affected by South-Eastern Mediterranean, mainly Balkan [Aegean], males.


See also:

31 comments:

  1. The total absence of Y-DNA I and very low percentage of mtDNA U5, the most common
    Epipaleolithic uniparental markers in ancient DNA is quite notable. Moreover, even if H is present in the Epipaleolithic in Southern and/or Southwestern Europe (a hypothesis which has credible evidence) the low levels of it in this sample are likewise notable and are small enough to have a predominantly Neolithic provenance.

    The high levels of Y-DNA G and low levels of Y-DNA R, however, support a hypothesis, that Sicilians (not unlike the type case of the Sardinians) have more first wave Neolithic, and comparatively less of the second wave Neolithic/Copper Age migration of farmer/herders than is found in much of the rest of Europe.

    The significant presence of Y-DNA E-V13 and absence of other Y-DNA E clades confirms (and maybe even drives) the PCA observation of similarity to the men of the Aegean and Levant relative to other West Eurasians.

    ReplyDelete
    Replies
    1. What do you mean by "second wave"? The Aegean Bronze clearly influenced much of Italy and particularly the South, and then of course they have all that Greek-like genetics, consistent with historical colonization as well, so there is clearly a "second wave" and even a "third wave" affecting this area. Just maybe not the "one" you have in mind (whichever it is).

      Delete
    2. I was thinking late Neolithic/Copper Age when I wrote "second wave", but really, in context and after considering your comment, it would perhaps be more fair to draw lines of "first wave Neolithic", transitional population genetic mixes, and the modern population genetic mix which is largely in place by ca. 2500 years ago or perhaps a few centuries earlier.

      Delete
    3. I guess that you're probably thinking in Central Europe's genetic paleohistory, right? I would say that each region had its own dynamics and merely extrapolating Central Europe's data to all the subcontinent makes no sense.

      Delete
  2. This comment has been removed by the author.

    ReplyDelete
  3. What is going on here? The 2013 report about Italy's genetic sex bias clearly stated that most of Italy's YDNA pool was pre colonial ie was Neolithic and Bronze Age. These results here seem to be all over the place, and contradicting the previous.

    ReplyDelete
    Replies
    1. That's just normal: different studies by different authors reach to somewhat different conclusions. It's even more madness if you look at the molecular clock estimates, which for me mean absolutely nothing (I just don't trust them at all - unless they are done on full Y chromosome sequences, very rare, in which case I may consider recalibrating them but never take at face value).

      What matters is the raw data (haplogroup frequencies and such). Reach to your own conclusions critically on that - at least it's what I try to do.

      Delete
  4. That was my fault then, I didn't realize it was different authors. Makes sense. Honestly, I see more validity in the "sex bias" report by Boattini, and I recently discussed this with both him and Luiselli, who both confirm the 2013 report. To me, just my opinion, this latest report looks messy to say the least. Example, on one area it says the genetic make up on both mtdna and ydna lineages in the South and Sicily are "homogeneous", and then describes a "melting pot". Other things are also strange. Im mtdna J2a1 for example, of Tuscan Matrilineage. ALL sources TMRCA date the minor J2a1 clade to have arose in Europe roughly 16,500 ybp. However, this new report gives some new and bizarrely archaic date for J2a clades in the South.

    ReplyDelete
    Replies
    1. Just for reference I added a "see also" section at the end of the entry with links to the latest studies on Italian genetics discussed in this blog. It's always good to contrast.

      TRMCA estimates (among other distorting factors, which are many) are affected by the local diversity. Say that in a given region there is presence of X1 and X2 and they measure the age of X as a whole, it will always produce an age older than that of X1 and X2 separately (unless STRs are by chance similar in both clades maybe). I don't think you can really estimate local ages but, at best, only haplogroup ages without any geographic constriction.

      There are many other issues: choice of method, calibration point (all scholarly methods heavily underestimate both the OoA age and the Pan-Homo split age), reliability of the set of markers used, etc. As I said, I don't take age estimates even half seriously: for me they are blank spaces, totally meaningless.

      Delete
  5. I'm pretty knowledgeable about this subject, but not on all the details. Thank you for the explanation. I comprehend for the most part. Another example of what you're describing would be my Ydna. My Paternal side is from Yorkshire, England, and Im the minor Haplogroup I2a1c (P37.2, L233), and is found at a low frequency (0.3%) in England and in Germany. This is refered to as the "Chauci" Subclade (the Chauci being an ancient Germanic tribe on the Continental North Sea coast), as the general demographic and TMRCA consensus is that its origin was the NW of what is now Germany. L233 is therefore generally held to represent the fewer I2a1s who settled NW Germania. As such, its presence in Britain is called an "Anglo Saxon" Subclade, since, given this scenario, it would have arrived with the Dark Ages Germanic settlement as a minority, lost in the dominant R1bs (Germanic Subclade) and I1s that comprised the bulk of those population movements. Nevertheless, you can still find, albeit rarely, source proposals that it was a possible "Isles" (Ancient British/Pre Roman) Haplogroup in Britain.

    ReplyDelete
  6. Cosenza is not in Sicily it is in Calabria. You wrote this above. ''while other areas of Sicily (Lecce, Cosenza, Enna)''.

    ReplyDelete
    Replies
    1. Alright, I'll edit that. Thanks.

      Delete
  7. I have a question if I may so do this here. Other genetic testing I did by far did not give any indicator of that, but on GEDmatch some people are telling me that supposedly they see an "inflated" "Aegean" / Greece-Anatolia contribution in my DNA. Others have said that my results look perfectly normal for a half English/half Italian, showing no "inflation". I know that the "West Asian", "Middle Eastern", and "East Mediterranean" readings are showing up because of Prehistoric (most likely Neolithic) Admixture, which is normal, but those who are telling me of an "inflation" are meaning that my results give reason to believe I had significant Post-Neolithic "Aegean" i.e. colonial Greek, Admixture. Again, if this is appropriate, I ask opinions please. # 174701

    ReplyDelete
    Replies
    1. I don't know the details of the test (compared pops. and such) but in general Southern Italians and Aegeans are not that different.

      In fact some recent studies on ancient DNA (Lazaridis 2013 specifically, see here) appears to show that Sicilians and Maltese are abnormally too "West Asian", even more than "allowed" by admixture of ancient Neolithic farmers (form the Alps I must say) and ancient Western European hunter-gatherers. Of course it may be something very old but it is also anomalous regarding Europe, because not even Greeks show that "poor fit" with the triple admixture model, although Ashkenazi Jews do.

      It's just one study and the reference populations are not directly related to Sicily but it does suggest that there's something different about Sicilians (and maybe other South Italians).

      Notice please that the main Greek colonial power was Phocaea (near Smyrna/Izmir in Asia Minor) and also that Sicily had some important Phoenician colonization, as well as becoming later the main Roman "plantation colony", what implied massive import of slaves from all around the Mediterranean. I don't see any obvious legacy in haploid DNA of this last element but at least the two others should have influenced Sicily in quite significant ways, more so as the Aegean influences are as old as Chalcolithic and probably also Mycenaean age (not only the Magna Graecia period allows for Aegean inflows).

      Delete
  8. I see what you mean. My East Mediterranean scores are roughly 10-15%, and West Asian and Middle Eastern are minor, so I don't think, at least as far as I know, this "inflation" scenario for me being more "Aegean" is correct. My Italian side was from Tuscany and Abruzzo (no Post-Neolithic eastern Admixture) and Campania. Campania would be the most reasonable candidate if there was this "colonial" or Post Neolithic scenario at play with my DNA. Anyway, this "East Mediterranean" doesn't necessarily mean "Aegean" or "Near Eastern" in GEDmatch terms as far as I've seen. The map used for that reference seems to scope from the Middle East to around Corsica (very broad), so perhaps that is the only reason why.

    ReplyDelete
    Replies
    1. "My Italian side was from Tuscany and Abruzzo ... and Campania".

      Ah, I though you meant Sicilian (because of context and also because it's the typical source of emigrants to North America.

      Although mainland Italy (including Tuscany) is a good fit for the Lazaridis model, that does not mean that there was no Aegean influences in the peninsula: there were some in the Chalcolithic (not limited to the South at all but also Central Italy - we don't know the genetic impact, just some cultural influence apparent in the archaeological record) and many suspect Etruscans to have been formed in a later Aegean migration which would have formed the Villanova culture (this was once apparently supported by ancient mtDNA but has been contested later on). Another support for this is the persistence of an Etruscan-related language in the island of Lemnos, near ancient Troy, in historical times and, indirectly, the Roman legend about the foundation of the city by Aeneas, which might have been borrowed from Etruscans (?), who ruled the city for some time and strongly influenced Roman techno-cultural progress (basically Roman architecture is Etruscan, as is their alphabet, etc.)

      I personally do lean for an Eastern origin of the Etruscan elites, for the reasons mentioned above but also because some of their words do look oriental: lukumon (king) resembles Sumerian lugal (king), Nept (Neptune) resembles the Egyptian goddess of the Nile Nephtis, etc.; also their hair-styles initially were totally Minoan-like, as was to some extent their art; and finally modern Tuscans generally deviate a bit eastwards when compared with other Italians. Their arrival c. 1300 BCE is coincident with the period of Mycenaean naval expansion (c. 1500 into Crete and also influences as far west as Iberia, c. 1100 "Sea Peoples" rampage: destruction of Troy, Ugarit, colonization of Cyprus...) So it seems reasonable to imagine that they were pushed westwards by the Greek Bronze Age expansion somehow. The legend of Aeneas may have some real base after all.

      "Anyway, this "East Mediterranean" doesn't necessarily mean "Aegean" or "Near Eastern" in GEDmatch terms as far as I've seen. The map used for that reference seems to scope from the Middle East to around Corsica (very broad), so perhaps that is the only reason why".

      That looks like a very reasonable explanation indeed. If Italy is part of that "Eastern Mediterranean" element, there you have it.

      Delete
  9. I am half Sicilian, and the region in Sicily where my paternal family came from show fair skinned, steel blue eyes and dishwater blonde hair. My family is from central Sicily, the Enna Province, from the town of Leonforte. Nearby towns of Nicosia, Sperlinga, Piazza Armerina and Aidone speak a Piemontese dialect brought to these towns 500 year ago from Pienmontese immigrants to Sicily. The Sicilian people are an admixture of Northern European, Central European, Greek, Arabic, Spanish, French, Norman, and even a bit Germanic. But, there is NO negro DNA found in the Sicilian people.

    ReplyDelete
    Replies
    1. All Europeans have "Negro" DNA, don't be naive nor prejudiced. Africa and Europe are too close to each other and genetic flow in both directions is clearly demonstrated, particularly in the Neolithic.

      Delete
    2. Peter I don't know why you speak about Arabic admixture since the Genetics show that the J1 in Sicily and South Italy founded in that study is older than the short muslim dominion.

      Look at that: "However, for the types found in Sicily and in the South it was calculated age of 3261 years ± 1345 corresponding to the end of the Bronze Age."

      Delete
  10. Agrigento was not a Phoenician city to begin with, since that people only founded two small emporiums in the western tip of the island (Palermo and Mozia), unlike tens of Indigenous Italic settlements and at least 20 Greek colonies.
    Agrigento was founded by colonists from Crete and Rhodes, here there aren't samples from those two Greek islands for a comparison but I have seen they have high J Y-Chromosome and some R1b U152.

    ReplyDelete
  11. Agrigento was not a Phoenician city, that people only founded two emporiums in the western part of the island (Palermo and Mothia).
    Agrigento was one of the most important ancient Greek city of the world, it was called Akragas and founded by Cretans and Rhodians.
    Unfortunately there aren't samples from those two Greek islands but I'm pretty sure they have genetic overlap for sure, at least I know the Greek islands tend to have a significant portion of Y-Chromosome J.
    About the slaves: Roman empire carried slaves from all empire in all of the empire (including Spain of course) basically not exclusively in South Italy as you said.

    ReplyDelete
    Replies
    1. Well, the fact is that Agrigento looks rather "Lebanese" (would be ancient Phoenician), unlike other Sicilian areas, which look more "Turkish" (would be ancient Greek). I can't help it if the genetic facts seem to contradict historical "knowledge". However...

      Maybe it has to do with the fact that Agrigento was settled from Gela, which in turn has origins not in Phocaea but in Crete and Rhodes, which may have a different genetic makeup (certainly Crete is different from the rest of Greece). The tendency towards Lebanon may in fact mask that Cretan origin (Cretans have lots of J2, while mainland Greeks do not but Lebanese do too): high J2 and low E1b would look Lebanese but also Cretan, so maybe that's the answer.

      "About the slaves: Roman empire carried slaves from all empire in all of the empire (including Spain of course) basically not exclusively in South Italy as you said".

      AFAIK that's not really correct, Italy and Sicily had a much larger slave population than any other part of the empire (although maybe there were other exceptions). In many areas traditional farming (often with free farmers, at least in the West) persisted as dominant until the late Empire's feudalization that attempted to enslave all free farmers (what in turn triggered the Bagaudae revolts and the collapse of the Western Empire, already very weakened by the lost of the Eastern colonies to the new Christian Neo-Hellenistic order).

      Delete
    2. Without samples from Crete and Rhodes we can't speak about Punic influences in Agrigento which aren't hystorically confirmed.
      I have seen tens of studies about the Greeks and there is an huge differences between mainland Greeks and islanders.
      Usually Cretas are more near eastern shift than mainland Greeks and that can play a role in the fact that Agrigento (ancient Akragas) is a bit different from the rest of Sicily.
      About Catania in Northern Italian direction (47% of R1b in that study) I think it has to do with ancient Sicels and a strong Roman settlement.
      About the slaves: it depends, some parts of Italy (including Sicily) had latifundium but slaves can not only exclusive from all the Mediterranean but there were many Gauls and Germans as well.
      But i don't think they left a significant influences among the Italian gene-pool.

      Delete
    3. The tendency towards Lebanon is only in the Y-DNA graph, there is no such thing (rather the opposite) in the mtDNA graph. Therefore we should focus this part of the discussion on Y-DNA.

      And we are lucky because there was a quite detailed Y-DNA review of Greece years ago and shows clearly that Cretans are distinct from mainland Greeks, with much less E1b and much more J2. There was no data for Rhodes. See: http://leherensuge.blogspot.com/2008/05/greek-y-dna-review-at-dienekes.html

      So that's probably it because you are right that Agrigento was a Greek and not a Phoenician colony. Makes sense.

      As for Catania, I can't say, because such high frequencies of R1b are not common neither in Greece nor most of Italy, including Lazio.

      About slaves I agree that no obvious influence is apparent. Their life was probably quite brutal and short, especially in the latifundia and mines (home slaves on the other hand had better lives usually but those were not so important numerically and often were of Italian origin).

      Delete
    4. Yeah I know about Crete.
      There is also some U152 in that island.
      About Catania: I have read the complete paper and it's surprising the high U106 (around 12%) and it's very surprising indeed.

      Delete
  12. Not at all, Sicilians are primarily an Italian population with Greek input in the coasts.
    The rest, especially Spanish, North Europeans, French and Arabs left a very small input in the Sicilian gene-pool.
    Frederick II removed, exterminated and expelled all the muslims from the island and most of them were just local Sicilians converted to islam.
    And other populations you have mentioned didn't setted in the island en masse.
    The Piedmontese who settled your part came from 1000 years ago for replaced the muslims who were expelled by Roger I not just 500 years ago.

    ReplyDelete
    Replies
    1. Well, all that is something we can't know for sure until we look at the genetics, right? That's what we're trying to do here, not just parroting the well known historical narrative. Sometimes the real people's history is not well reflected or at all in know history, you know. Especially ancient historians often did not feel the need to talk of commoners, only about the elites - but elites are by definition exceptional.

      Not sure what's up with the Piedmontese, sounds intriguing.

      Delete
    2. My last reply was for the guy Peter who has spoken about northern Italian settlements in Sicily in the middle ages.
      Anyway the normans used mainlander Italians to repopoulate part of Sicily, but they were mostly from mainland South Italy rather than northern Italians who settled in specific areas.

      Delete
    3. My mtn Dna is j1 c1. Am 100%sicilian, and I now the names of my ancestor
      Since 1740 . I know remenello remains find neat the Po river had the same
      Mtn dna I'm from agrigento area. Very intrigued about j1c1... Is that easily found in Sicily?

      Delete
    4. What we know from this study is that J1 (in general) makes up about 5% of Sicilian genetics. So in principle it should not be a rare lineage.

      Something that is plausible about J1 is that it probably arrived in connection with Phoenicians or Arabs. But we cannot say with 100% certainty, as some J1 also exists in the NE Mediterranean, so it could also be Greek, etc. (but less likely).

      I've been as of late caressing the idea of ancient Siculi/Sicels being a Semitic people, the Shekelesh of the Sea Peoples' narrations, who may have arrived in Italy together with the Etruscans (Teresh = Tyrsenians) and maybe "invited" or "incited" by the Sardinians (Shardana or Sherden) to fight against the invading Indoeuropeans (Italics, proto-Latins included) who were then advancing from the North. After being defeated in mainland Italy, probably in Latium itself, they settled Sicily and Calabria. This could be another way for J1 lineages to have arrived to Sicily and Italy in general but it's a tad more hypothetical, as we're dealing with reconstructed proto-history.

      In any case my theory is based on the fact that the Shekelesh are described as "circumcised" (as Israelites and other Semites) but not like Teresh (Trojans/Dardani quite plausibly) and other Euro-Asian peoples of the North. The name Shekelesh might mean "mercenaries", as "shekhel" is an old weight measure and later coin in the West Asian area.

      Delete
    5. PD- About my theory on Shekelesh, "si no é vero, é ben trovatto". Non lo so.

      Delete

Please, be reasonably respectful when making comments. I do not tolerate in particular sexism, racism nor homophobia. Personal attacks, manipulation and trolling are also very much unwelcome here.The author reserves the right to delete any abusive comment.

Preliminary comment moderation is... OFF (keep it that way, please)