search this blog

Wednesday, November 11, 2015

Genetic substructures in Italy


The main text is behind a pay wall, but the images and supp info freely available here and here. Abstract:

Recent scientific literature has highlighted the relevance of population genetic studies both for disease association mapping in admixed populations and for understanding the history of human migrations. Deeper insight into the history of the Italian population is critical for understanding the peopling of Europe. Because of its crucial position at the centre of the Mediterranean basin, the Italian peninsula has experienced a complex history of colonization and migration whose genetic signatures are still present in contemporary Italians. In this study, we investigated genomic variation in the Italian population using 2.5 million single-nucleotide polymorphisms in a sample of more than 300 unrelated Italian subjects with well-defined geographical origins. We combined several analytical approaches to interpret genome-wide data on 1272 individuals from European, Middle Eastern, and North African populations. We detected three major ancestral components contributing different proportions across the Italian peninsula, and signatures of continuous gene flow within Italy, which have produced remarkable genetic variability among contemporary Italians. In addition, we have extracted novel details about the Italian population’s ancestry, identifying the genetic signatures of major historical events in Europe and the Mediterranean basin from the Neolithic (e.g., peopling of Sardinia) to recent times (e.g., ‘barbarian invasion’ of Northern and Central Italy). These results are valuable for further genetic, epidemiological and forensic studies in Italy and in Europe.

Citation...

Fiorito et al., The Italian genome reflects the history of Europe and the Mediterranean basin, European Journal of Human Genetics advance online publication 11 November 2015; doi: 10.1038/ejhg.2015.233

29 comments:

Simon_W said...

The basics of this have been known for a while, thanks to the Dodecad project and related efforts, but as a 1/4 Italian I find it very interesting to learn more about the details of the geographical structuring, so thanks for posting.

My special interest pertains to the relative position of the Emilia-Romagna. As the supplementary tables show, all EMR samples were from the province of Ferrara, that's not exactly where my ancestry is from, but still good enough for the Emilia-Romagna in general. The EMR samples cluster with the other North Italians in the Italian PCA, and they're particularly close to the Ligurians, the other North Italian sample from south of the Po, though a couple of EMR individuals is very close to the Tuscan epicenter. What I inherited from my Italian grandfather is quite Sardinian-shifted in the West Eurasian PCA. And in this Italian PCA, closest, or rather least remote from Sardinians are two Lombard individuals.

The ADMIXTURE analysis within Italy at K=3 makes complete sense: There is a blue Sardinian component = EEF, a red Calabrian centered component = East Mediterranean-Caucasus-related input, and a green Val d' Aosta centered component = central European shifted, but still rather Mediterranean, judging from the Iberian-like position of VDA in the West Eurasian&North African PCA. The red component is strong in the south and in Sicily and gets progressively weaker towards the north. Mark well: Tuscany doesn't score particularly high in this component. But this has been known before. The Sardinian component is everywhere weak except on Sardinia. But of course the red and green components have eaten up some EEF ancestry.

The West Eurasian&North African PCAs resemble the ones posted by David.
Northern Italy is inbetween Iberia and central Italy, and southern Italy is inbetween central Italy and Armenia&Cyprus.

The West Eurasian&North African ADMIXTURE analysis is most interesting at K=5. There is some slight African (Mozabite centered) admixture in Sardinia and Southern Italy, but this is negligible elsewhere. There is also some slight Saudi/Bedouin centered admixture in the south, which gets even slighter towards the north. The remaining three components behave as could be expected: The Druze centered „West Asian“ red component is strongest in the south and gets progressively weaker towards the north. The Sardinian&Basque centered „Atlantic-Med“ component behaves in the opposite manner, and the green East European component as well.

truth said...

@Simon
The K3 admixture is intra-Italian, blue is Sardinian, Green is Aostan, and red is Calabrian, you can see how the other samples relate to each other with these components. For example LOmbards would be mostly like aostans more a bit more southern shifted (red component) and west-shifted (a bit of Sardinian compnent), that also correlates with PCA.

andrew said...

One point I recall from some several year old studies of modern Italian population genetics is that there are some relict isolate populations, basically in the far Northeast of the country the highlands between Venice and the rest of Italy more or less, that are very distinctive relative to the rest of Italy or even to the rest of Europe in general.

These populations, while small, could potentially have an importance to archaeo-genetics in Europe comparable, for example, to that of Sardinians, Basque populations, and the Saami in piecing together what the European genetic landscape looked like prior to the mass demic migration from the Steppe ca. the Bronze Age, which was the final major step in establishing the modern population genetic makeup of most of Europe (with a few exceptions mostly associated with legacies of the Islamic Empire in Southern Iberia, Malta and parts of the Balkans and modern era colonial immigration).

I can't recall for the life of me the name of this ethnic enclave, but it would be interesting to see if that population is included or discussed in the current study.

andrew said...

This study http://www.readcube.com/articles/10.1002%2Fajhb.22677?r3_referer=wol&tracking_action=preview_click&show_checkout=1&purchase_referrer=onlinelibrary.wiley.com&purchase_site_license=LICENSE_DENIED seems to be the one I'm thinking about even though the date of publication seems much more recent that I recalled.

Krefter said...

IBD, ADMIXTURE, and PCA are all telling the same story......

Italians are mostly EEF with North European, West Asian, and minor NW African ancestry in South Italy+Sardinia.

The IBD sharing between North Europe and North Italy is very high. IMO, this is because of Bronze age North European or any combination of EEF+Steppe admixture. It could be from later admixture events. If the upcoming Estruscan genome shows the same high IBD it must be older.

IBD doesn't suggest recent West Asian ancestry at all. But ADMIXTURE does.

Krefter said...

IBD stats.

https://docs.google.com/spreadsheets/d/1FQOJVWKiN_Y4-IoUUdPIjqfzdN-zH9R-9C_vwcbciCw/edit#gid=286859889

Joshua Lipson said...

We need Jewish-Italian IBD analysis on the regional level.

Krefter said...

@Davidski and others,

Can you run these IBD stats.....

>Greece, Albania, Serbia with all Italian genomes.
>A few Spanish with Syrian and Palestinian.
>A few British with Syrian and Palestinian.
>A few Spanish with British.
>Remedello with all Italian genomes.

Maju said...

A complementary also new material on Italy: http://www.biomedcentral.com/1471-2156/16/132 - it deals with purported selection in some alleles but the general analysis produced is the same.

@Simon: I'm also 1/4 Italian, from the same area as you, although in my case I can track the oldest known male ancestor some 500 years ago right in Ferrara, another known female ancestor is from Rome though. Considering what we know about how people relate to each other (i.e. all Europeans are relatives among each other from, at most, 500 years ago), I guess I'll have to consider you a "cousin" of sorts, right?

Maju said...

A detail that has caught my attention is the position of Aostans in the all-Europe PCA, although it varies with or without Ötzi the Iceman. Without him, Iberians are drifted East relative to North Italians and Aostans, who appear more Basque-like than Iberians (tending to France in the case of Aostans and to Sardinia maybe in the case of North Italians). With Ötzi instead the Iberians are relocated and do not anymore appear "more Eastern" than North Italians. In any case there is a cluster of N.It., Aos and Ibe. Any thoughts?

andrew said...

The Otzi effect is weird. It could be that there are several moderately similar strains of EEF, perhaps one derived from CP and one derived from LBK, and a third which is extinct and represented by Otzi, who is closer to LBK than to CP. In the absence of Otzi, the statistics may be trying to use an LBK strain of EEF as a proxy and pulling too far to the east. With Otzi to provide the intermediate strain of EEF, the slightly mismatched strain of EEF may resolve.

Simon_W said...

@ truth

Yes, that was clear to me, I was referring to „the ADMIXTURE analysis within Italy“, by this meaning „intra-Italian“.

@ andrew

No, there are no samples from northeastern Italy included, not north of EMR at any rate.

@ Maju

I almost fell of my chair! I always believed you were 100% Basque!
Yep, given the high endogamy on a regional level, we probably share some not too distant ancestors. I didn't do any genealogical research, to me it sufficed to know that my grandfather's parents were from the province of Forli-Cesena. He believed to be descended from a forebear from Reggio Emilia, but I'm not sure if that's correct. He also mentioned the usual Roman ancestor.

Simon_W said...


IMO the central European/northern admixture in Italy can't be reduced to „Barbarian“ invasions. I suppose this term refers to the Celts and migration age Germanic influence. IMO the admixture is most of all from the Italics themselves, who, as it seems, were from Bronze Age central Europe (including Hungary). The international ADMIXTURE run shows the Chuvash centered „northern“ component to be hardly weaker in central and southern Italy than in the north.

And IMO the East Med/Caucasus-West Asian admixture can't be reduced to the Greek colonisation after 800 BC. Because that admixture is also very strong in Apulia, which however was largely bypassed by the Greek colonists. Rather, it's probably related with what Harrison and Heyd wrote in the recently linked 2007 paper: That the central Mediterranean (with Sicily, Apulia and Malta) between 2600 – 2200 BC was drawn into the wider Near Eastern exchange network centered on the high cultures of Mesopotamia. And the eastern contacts continued in the second millennium BC.

Simon_W said...

On a more personal note, I've changed my mind, I don't think what I inherited from my Italian grandfather can be very EEF-shifted; my own somewhat EEF-shifted position relative to my other grandparents is probably the result of recombination magic, I guess I inherited a little more Iberian Bell Beaker ancestry from my central European grandmother than of her more northern stuff.

Ponto said...

Regarding Malta, the present day Maltese derive from the colonization of the islands from Sicily in the 11th century. The islands were depopulated by the Muslims in the 9th century. So it does not matter what transpired in Malta prior to the 11th century, 1000 years ago.

What I find interesting is that Calabrian Italians are south and east of the Sicilians, meaning that Calabrian Italians have more IBD sharing with North Africans and Near Easterners than Sicilians who were under the control of North African muslims for centuries and the Calabrian Italians were not. There is an anomaly that needs explaining.

André de Vasconcelos said...

@Ponto

That's not really unheard of. In Iberia there is a West-East divide (if you can even call it that) in which Westerners are very slightly more related towards NW Africa than Easterners. But historically the West was under muslim control for a much shorter period of time (be it Umayyad, the local Emirate or just a Taifa). Presumably most of this ancestry dates back to before 711. Maybe the same happened in southern Italy.

Matt said...

@ Andrew: Friuli-Venezia, yeah, I remember posting about that earlier in the year, noting the similarity in shape to a PCA involving them to the Eurogenes standard PCA:

http://i.imgur.com/ZX3V1zu.jpg

Paper was "Genetic characterization of northeastern Italian population isolates in the context of broader European genetic diversity". The same samples were used again I think in this paper

http://biorxiv.org/content/biorxiv/early/2015/07/21/022947.full.pdf

I don't think there is likely to be anything too totally new with them, but the Resian isolates may be like the Bronze Age Hungarians or more extreme, in terms of the balance of EEF, WHG and Yamnaya components, while the Resian general population looks like it could be like the Basque. (Or they may just be drifted!)

I dropped Reich an email about all that, before the Yamnaya paper, but no response or anything. Shame they aren't open access for anyone to check.

...

On another note, kind of cool to see the West Eurasia PCA here, as without a large West Asian sample whose contrast to Sardinians dominates dimension two, the fairly large differences between North Africans / Levantines and Sardinians / SW Europe are more visible. So it gives a different view.

truth said...

David :
are these samples available ? I would like see them on your PCA like these ones :

http://imagizer.imageshack.us/a/img834/7859/c0zu.png

especially the Aostan sample i'm interested in.

Maju said...

@Simon: "I always believed you were 100% Basque!"

Nah, that's what some people imagine but actually I can only trace my father's family (100%) and a small part of my mother's family (my great-great-grandmother) to the Basque Country, all the rest are Spaniards (incl. La Rioja but with unclear further ancestry) and then there is my Italian (maternal) grandfather (whom, being aristocratic, could trace parts of his ancestry to the 16th century or even earlier: a knight from Ferrara via direct patrilineage, and Santa Francesca Romana, via another lineage). I'm largely a product of the forths and backs of Fascism in Europe (ancestors usually suck, at least in my experience, but we are not our ancestors: people belongs to where they live or at the very least to where they grew up).

So I guess I'm just 56% Basque by ancestry but 100% Basque by birth and heart. And that is what really matters.

Maju said...

@Ponto: I have this hypothesis that maybe the Siculi or Sicels were the Shekelesh of the Sea Peoples, who were a circumcised people and therefore most likely Semitic (the name could just mean "mercenaries" if related to the historical monetary/weight word "shekel"). That would explain to my eyes better the kind of hyper-Mediterranean slant we find among the usual Sicilian samples than the appeal to Phoenician or Muslim origins. There's no reason why the Shekelesh could not have extended into Calabria prior to Italic conquest (we just don't have enough data for the period, I fear), however it could be more easily supported if we knew of some sort of cline between West (non-Sicel) and East (Sicel) Sicilians. AFAIK no such research has ever been performed and Sicilian genetic homogeneity is assumed by default.

I found this in the Italian Wikipedia (myths regarding Calabrian origins):

secondo il mito greco, circa 850 anni prima della guerra di Troia, vi sarebbero dunque giunti Enotrio e Peucezio (riportato anche come Paucezio), di stirpe enotria e pelasgica, originari della Siria che, trovando il suolo molto fertile, chiamarono la regione "Ausonia" in ricordo dell'Ausonide, fertile zona della Siria.

In brief: according to a Greek legend, some 850 years before the Trojan War peoples original from Syria, the Enotri, established themselves in what is now Calabria and nearby areas, naming it Ausonia in memory of fertile parts of their Syrian motherland.

If you read the Enotri entry, they instead mentione that the settlement would correspond with the Iron age and that the Enotri split in three branches: the Itali and the Morgeti in Calabria and the Siculi in Eastern Sicily.

Other articles however consider them to be Italics of the Oscan branch but this may correspond to the usual tendency to oversimplify including any question marks in the mainline population that is most celebrated by historiography (as happens with Celts in Western Europe, etc.) or also to ethno-linguistic assimilation, which I consider plausible in the case of the Siculi, whose language looks anything but Semitic.

In any case I'd consider seriously the possibility of Sea Peoples' genetic influx from the Eastern Mediterranean: the Shardana (Sardinians) would simply return home after a century of mercenariate but may have brought with them some exotic allies, the Shekelesh Semites and the Teresh Aegeans (Etruscans), who conquered some lands and may have impacted the genetics of parts of the peninsula and Sicily significantly.

Food for though it any case.

Gaspar said...

To get the facts of association of the North-east italians to other italians, then check
Robino, C.; Ralf, A.; Pasino, S.; et al. (2015). "Development of an Italian RM Y-STR haplotype database: Results of the 2013 GEFI collaborative exercise". Forensic Science International: Genetics

Basically they only truly assimilate with people from Marche.

History of the Area states the raetic and their close cousins the Euganei ruled all these lands in the late bronze-age.
The Veneti ( eneti ) from Anatolia speaking a Palaic language
https://en.wikipedia.org/wiki/Palaic_language
occupied and absorbed the Euganei accepting the Euganei langauge ( similar to Raetic, camunic, lepontic scripts

Later mass migrations came from the eastern goths ( ostrogoths) who had previously conquered the sarmatians and absorbed them, settled in NE-Italy for over 250 years ..............

Maju said...

@Gaspar: Venetic appears to be a Western IE language with clear links to Italic and possibly Germanic and (the always controversial) Illyric: https://en.wikipedia.org/wiki/Venetic_language

So why do you claim that Venetic could be related to such a poorly attested language as Palaic? I don't make any sense of it.

Gaspar said...

@maju

You did not read what I said, the Veneti gave away their Palaic language accepting the indigenous language of the Euganei which we now call venetic. ( thats the gist of Italian historians on the Paleo-veneti )

Cato states the Euganei having 34 towns, we cannot imagine a huge venetic contingent arriving in NE-Italy..........it was a small migration.
Basically the Venetic people mostly comprise of the indigenous Euganei

Simon_W said...

@ Ponto

At least the Maltese are genetically distinct from the Sicilians and speak a Semitic language. So at any rate they're not the same as Sicilians. This doesn't contradict what you wrote, you may be right, I don't know.

@ Gaspar

Your source is about y-DNA. It tells us nothing about the autosomal affinity. In all likelihood autosomally, Venetians are closer to other North Italians than to central Italians.

@ Maju

My Italian grandfather also believes to be descended from the ancient nobility. Not sure if that has been thoroughly researched. At least it was claimed by the Istituto Araldico Genealogico Italiano, which however is a private commercial corporation.

An interesting approach you have; mine is rather different: First I want to find out which peoples I'm descended from, then I attach my heart onto them. :D But I guess they're both valid approaches.

Simon_W said...

@ Maju

Personally I'd rather equate the Shardana aka Sherden with the Serdi from Thracia than with the Sardinians. It would make sense given the Balkan epicenter of the Sea People migrations. The Serdi were located around a town called Sardon polis in Greek, showing that the first vowel varied between an e and an a, like in Sherden and Shardana. The horned helmets of the Shardana were certainly compatible with a Balkan origin.

And I prefer to derive the Sicels in Sicily from Bronze Age central Italy, as did Pliny the elder and Dionysius of Halicarnassus. Many ancient authors agreed that they were from mainland Italy, including Hellanicus of Lesbos, Philistus of Syracuse, Antiochus of Syracuse and the reputable Thucydides. According to the latter, the Sicels left mainland Italy about 300 years before the first Greek colonies were established in Sicily, that would mean about 1030 BC. Now, archaeology shows that in the very south of Italy, in Calabria and northeastern Sicily the Protovillanovan cremation custom was gradually given up already in the Final Bronze Age II, i.e. in the 11th century BC. The respective facies is called Ausonio II. And this subsequently spread in Sicily. To me it strongly looks like the bearers were the Sicels.

Regarding the internal genetic structuring of Sicily, I recommend the MDLP K23b analysis. It included several Sicilian samples: Sicilian_West, Sicilian_Center, Sicilian_East, Sicilian_Agrigento, Sicilian_Siracuse and Sicilian_Trapani. This is the spreadsheet:
https://drive.google.com/file/d/0B6n7iMc2P-yQUmFiVm9PTGMteG8/edit

The Near Eastern component is strongest in eastern Sicily and weakest in central Sicily.

Maju said...

@Simon: I also tended to think that the Shardana or Sherden would be just another Eastern Mediterranean people but I was persuaded by the issue of horned helmets, which happen to be documented in bronze figurines of Sardinian warriors of the Nuraghe period. The Shardana are also unlike all the rest of the Sea Peoples: they appear the first of all, c. 1278 BCE, and stay put all the time (because they became mercenaries for the Egyptians, who garrisoned them in the Levant and even used them as palace guards). The other Sea Peoples instead belong to two other very specific attacks much later:

c. 1208: the Nine Bows coalition was formed by Hittites, Libyans (Libu), Israelis (first mention ever), Meswesh (Tamazigh, i.e. Western Berbers from near Tunisia, known historically as Mazyes, Massagetes), Teresh (Tauresi = Tyrsenians), Lukka (Lycians), Ekwesh and Shekelesh (circumcised, probably other Semitic groups) and the Shardana (who were obviously turning against their Egyptian overlords). In brief: this is a Hittite attack along with a number of allies from the Aegean (Lukka, Teresh), the Levant (Shardana and the three Semitic groups) and North Africa (Libu and Meswesh).

c. 1175: attack against Ugarit and failed march south against Egypt. This was surely led by the Denejen (Danaoi = Achaeans = Greeks), who continued with their rampage from Troy to Cyprus (well documented archaeologically) and Syria. The Ugarit tables suggest that the war had started further West, where fleets and armies were deployed (and maybe already defeated) and was synchronous with the fall of Hattussa (to unclear foes), which asks for help to a powerless Ugarit. The Greeks brought with them the Peleset (Pelasgians?, certainly the later Philistines), Tjekker (Teucrians = Dardanians), the Shekelesh (Semites) and a mysterious people known as Weshesh.

c. 1178: failed attack against Egypt from the sea by the same Greek-led coalition but enlarged with the Shardana and the Teresh.

My hypothesis is that after this defeat (or maybe a few years later, after the final collapse of Egypt to the Meswesh Berbers, who became overlords of the Delta) the Shardana, Teresh and Shekelesh headed to Italy. Berber logistical support is plausible but uncertain.

Maju said...

@Simon: regarding the spreadsheet, I cannot download it (no downloading permissions given) and I don't see the column names in the view, so I will have to limit myself to take notice of what you say as synthesis: "The Near Eastern component is strongest in eastern Sicily and weakest in central Sicily". I presume that it means that West Sicily is intermediate, what may be attributed to Phoenician colonization.

Re. "Many ancient authors agreed that they were from mainland Italy"... should not be contradictory with hypothesis. It's perfectly possible that they conquered first parts of mainland Italy, close to those controlled by their Etruscan allies, but were driven out by the Italics. That process would also explain better that their language does not appear Semitic and may even have some IE relations (uncertain).

It may also be worth mentioning that Eastern Sicily was an important commercial connection of the Iberian late Bronze (around 1000 BCE) aka Atlantic Bronze. They imported elbow fibulae from Cyprus, which were also popular in the Iberian Plateau and the Etruscan countries (both Tuscany and Emilia), as well as some lesser scatter in Southern France and other parts of Italy. These imports are considered precursors of the Phoenician colonization.

Gaspar said...

@simon_w

correct , Venetians are closer to other north-italians because the bulk of Venetian ancestry comes from the indigenous Euganei people who where in Veneto and Friuli in the late bronze-age.
The Euganei hills ( colle) are name after these indigenous people
https://en.wikipedia.org/wiki/Euganean_Hills

The veneti part came from North Anatolian the Palaic area as per the conclusion and tests done by NatGeno between 2005 - 2010

But this paper has zero samples from anybody on the adriatic sea coast.
It's like the testers though it would askew the results too far.....or maybe they just wanted to find a west-atlantic affiliation

sabrina run said...

Italy, according to its history, is a country with diverse population genomics. I think to research on its population genomics is a very interesting and useful project. cp wants to offer help for it.