February 08, 2013

Etruscan mtDNA origins (Ghirotto et al. 2013)

From the paper:
A model of genealogical continuity across 2,500 years thus proved to best fit the observed data for Volterra, and especially Casentino, but not for another community dwelling in an area also rich with Etruscan archaeological remains (Murlo), nor (as expected) for the bulk of the current Tuscan population, here represented by a forensic sample of the inhabitants of Florence.
As for the second question, the IM analysis shows that indeed there might have been a genealogical link between modern Tuscans and the inhabitants of what Herodotus considered the Etruscans’ homeland, Western Anatolia. However, even under the unrealistic assumption of complete reciprocal isolation for millennia, the likely separation of the Tuscan and Anatolian gene pools must be placed long before the onset of the Etruscan culture, at least in Neolithic times; if isolation was incomplete, the estimated separation must be placed further back in time. Consistent with this view is the observation that Etruscan and Neolithic mtDNAs are close to each other in the two-dimensional plot of Figure S4C; however, a formal test would be necessary to draw firm conclusions from the simple observation of a genetic similarity. Separation times were very close when estimated both using a sample from Western Anatolia, and an expanded sample including individuals from much of Anatolia, and so the choice of the Anatolian population does not seem to affect the results of this analysis.
As always with estimates in years, the choice of mutation rate may affect results, but I am reasonably confident that this particular result does not depend on such issues. From the paper:
For these tests we chose the mutation rate (μ) estimated from the data in the previous ABC analyses (very close to the figure accounting for the time-dependency of the mitochondrial molecular clock [13], μ = 0.003). Tests were also run using the value incorporating a correction for the effects of purifying selection [23] (μ = 0.0014), always finding that it results in a further increase of the estimated separation times (Figure S7B). Only assuming very high mutation rates, at least twice as large as estimated in Henn et al. [13], was it possible to obtain separation times less than 5,000 years (Figure S7B). With both Anatolian samples, any degree of gene flow after separation between the ancestors of Tuscans and Anatolians resulted in more remote separation times.
A couple of observations:

If Etruscans did originate in Anatolia then presumably the historical Etruscans were not descended entirely from them but from a mixture of pre-Etruscans with the incoming population. So, it would seem that the inferred dates are incompatible with a folk migration model of Etruscan origins, but not necessarily with a model that accommodates admixture (e.g., initial mtDNA gene pool separation c. 8,000 years ago with the onset of the Neolithic + later admixture during the Bronze Age). On the other hand, the close similarity between Etruscan and Central European Neolithic mtDNA is a good argument for (mostly) continuity in this case.

That things did happen in Italy in the last 5,000 years can be inferred on the basis of the Iceman's genome. It will certainly be interesting to extract Y chromosomes and/or autosomal DNA from some of these Etruscan samples.

A different issue that may bias dates upwards is the occurrence of East Eurasian mtDNA in current Anatolian Turks. It is not clear by how much this would affect age estimates (this admixture is low, sub-10%, but from a population that split off from West Eurasians perhaps more than 40kya); it would nonetheless be useful to repeat the experiment after either (i) purging the Anatolian sample of lineages likely to have introgressed into the population in medieval times, or (ii) using a different West Asian sample other than that of Anatolian Turks.

In any case, it's great to finally have the genetic characterization of a historical European people, and hopefully more samples will follow both from Italy (at least from those who practiced inhumation) and elsewhere.

PLoS ONE 8(2): e55519. doi:10.1371/journal.pone.0055519

Origins and Evolution of the Etruscans’ mtDNA

Silvia Ghirotto et al.

The Etruscan culture is documented in Central Italy (current Tuscany and Northern Latium, formerly known as Etruria) between the 8th and the 1st century BC. Questions about the Etruscans’ origins and fate have been around for millennia. Herodotus and Livy regarded them as immigrants, respectively from Lydia, i.e. Western Anatolia, or from North of the Alps, whereas for Dionysius of Halicarnassus they were an autochthonous population [1]. Previous DNA studies, far from settling the issue, have raised further questions. The Etruscans’ mitochondrial DNAs (mtDNAs) appear similar, but seldom identical, to those currently observed in Tuscany [2], [3]. Assuming reasonable effects of genetic drift and mutation, these levels of resemblance proved incompatible with the notion that modern Tuscans are descended from Etruscan ancestors [4], [5]. Explanations for this result include the (extreme) possibility that the Etruscans became extinct, but also that their modern descendants are few and geographically dispersed, or that the ancient sample studied represents a small social elite rather than the entire population [4]. As for the Etruscans’ origins, ancient DNA is of little use, because pre-Etruscan dwellers of Central Italy, of the Villanovan culture, cremated their dead [1], and hence their genetic features are unknown. DNAs from modern humans and cattle in Tuscany show affinities with Near Eastern DNAs, which was interpreted as supporting Herodotus’ narrative [2], [6], but in these studies modern Tuscans were assumed to be descended from Etruscan ancestors, in contrast with ancient DNA evidence [5]. The claim that systematic errors in the Etruscan DNA sequences led to flawed genealogical inference [2], [7] is not supported by careful reanalysis of the data [8].



Anonymous said...

All modern Italians including Tuscans are largely affected by Near Eastern or Eastern Mediterranean immigration from the Neolithic farming cultures, as are many other Europeans. I am incredulous how that can support the Anatolian origins of the Etruscans as distinct from any immigrants from the Eastern Mediterranean in the last 10,000 years. Anyway what do modern peoples in Anatolia, Europe or parts of the Middle East tell us about modern peoples and their ancestry? Nothing really except suppositions.

mooreisbetter said...

More pseudo-science from scientists determined to insert huge wedges in the Italian population.

Ask yourself this: what if a population of those few locales had an infkux of Anatolian slaves durinng Roman times, followed by isolation and drift during the Dark Ages? False positive.

Why are those 2-3 sites always picked on as being the most Etruscan today? There is no evidence that those cities they always use have somehow been kept pure over the ages. False positive.

Look up the history of Volterra. The Romans killed ALL the inhabitants. And another siege did so again later in history. If a twice wiped out/ twice repopulated city is their best proxy for an ancient populalation, have these authors read their history? (No). Or do they have an agenda? (Yes.)

Why were no other isolated Etruscan small towns sampled? Because the mtDNA there shows continuity.

Is it laughable that Florence would be even mentioned in the study? Since when do large cities with much in-migration mean anything when making ancient inferences?


andrew said...

Another hypothesis from antiquity traceable to Pliny the Younger in his Natural History (79 CE) where he wrote that "adjoining these (the [Alpine] Noricans) are the Raeti and Vindelici. All are divided into a number of states. The Raeti are believed to be people of Tuscan race driven out by the Gauls; their leader was named Raetus." The Etruscans, of course, are predominantly associated with the Tuscan region, so the "Tuscan race" would have referred to them at the time, in contrast to the Italic peoples like the Romans who had migrated to Italy more recently.

Pliny's link of the Alpine Raeti and the Etruscans is confirmed linguistically by Helmut Rix, and also archaeologically. Villanovan material culture migrated from the Alpine area to Tuscany around the time of Bronze Age collapse. (N.B. modern Swiss Rhaetic derived from Latin, and Iron Age Swiss Rhaetic languages, are completely different languages that happen to share the same geographically derived name.) Both the Rhaetic retreat to the mountains and the secondary Etruscan migration to Tuscany were probably driven by the "push" of early proto-Italic and Celtic populations. Indeed, one of the likely reasons that the Etruscans survived as a distinct culture longer than any other the other non-Indo-European cultures of Southern Europe (except the Basque) is that they adopted culturally many of the innovations of the Indo-European Urnfield culture and thus could compete with it. The theory of Herodotus about Western Anatolian origins was rejected by Greek Historian Dionysius of Halicarnassus for a variety of solid linguistic and religious culture grounds even at the time it was offered. A Bronze Age migration from Western Anatolian also suffers from the fact that Western Anatolia would have been a linguistically Indo-European area (not consistent with the non-Indo-European Etruscan language) for most of the Bronze Age.

This mtDNA study confirms once again that the Etruscans were surely not derived substantially from Upper Paleolithic indigeneous hunter-gatherer populations of Europe, whose mtDNA was dominated by mtDNA haplogroup U4 and U5. And, ancient DNA evidence also suggest that Etruscans were intrusive to Italy in the archaeologically supported early Iron Age period. Specifically "Genetic distances and sequence comparisons show closer evolutionary relationships with the eastern Mediterranean shores for the Etruscans than for modern Italian populations. All mitochondrial lineages observed among the Etruscans appear typically European or West Asian, but only a few haplotypes were found to have an exact match in a modern mitochondrial database, raising new questions about the Etruscans’ fate after their assimilation into the Roman state." Areas of historical Etruscan occupation also have a relatively high concentration of y-haplogroup G, which is characteristic of first wave Neolithic populations.

Etuscans are thus derived from people who arrived as part of a folk migration in the first wave Southern European Neolithic Cardium Ware culture that included Southern France, Sardinia and all of the territory attributed to the hypothetical Tyresian language family to which the Etruscan and Rhaetic languages belong. This culture, in turn, was derived from what is now Syria and Southern Central Anatolia rather than Western Anatolia, although both the donor and receiving regions have seen massive demographic upheaval in the intervening 7500 years (a time frame consistent with the latest mtDNA analysis). Other populations of modern Europe with significant Cardium Pottery ancestry probably include the Sardinians and the Galicians of NW Spain (but not the Basque whose genetic distinctiveness and ethnogenesis probably have roots in a second wave of folk migration by Bell Beaker peoples in the Copper and early Bronze Ages).

GailT said...

There is no mention of haplogroups in the paper, and it appears they did not attempt to assign the ancient DNA samples to haplogroups, so I looked at the 30 ancient mtDNA haplotypes listed in Table S1.

Seven of them appear to be U5, although they only identified the U5 defining mutation 16270 in two of the seven, and they seem to have missed several other mutations that should be present. For example, samples Hap4 and Hap5 are both U5a2a (based on the combination of 16114a and 16294). U5a2a should also have mutations at 16256, 16270 and 16526, but they miss all 3 of these in Hap4, while they only found 16256 in Hap5. So based on the U5 samples, they appear to have a high error rate of missed markers in their results.

Eight of the samples appear to be J, based on the mutation at 16126. Two of the samples might be H1b based on the mutation at 16356. I can't identify haplogroups for any of the remaining 13 samples. Six of the samples are CRS, but given that they missed 16270 in most of the apparent U5 samples, it is really impossible to guess what haplogroup the CRS samples might be.

Figure 3, the median joining network, has the haplotypes scrambled in a way that has no connection to their actual relationship in the phylgeographic tree. So that makes wonder if the rest of their analysis is meaningful.

Obviously they need to do additional sequencing on 13 of the 30 samples to identify their haplogroup. But the results we have so far seem to show the ancient Etruscans samples dominated by haplogroups J and U5.

Crimson Guard said...

Volterra is closest to Sicily in mtDNA:


Etruscan crania are closest to Sardinians:


Rob said...

1) Good work GailT !

2) There are widespread problems with "scholars" taking Herodotus' folk tales and folk etymologies at face value. Apparently, many historicans still fail to critique and contextualize primary 'historical' sources , like Herodotus.

Rob said...

(3) Further, it appears much of Europe beyond the Atlantic coast and Balkans was non-IE speaking even just prior to the Roman expansion; no matter what the "Kurganists" speculate about the direction of IE spread. The historico-linguistic facts do not sit with a spread from the north-east.

Anonymous said...


You forgot something. Please google "Lemnian language".

Ebizur said...

Crimson Guard wrote,

"Etruscan crania are closest to Sardinians"

Actually, this dendrogram's "modern period Sardinia" sample is closest to its "Neolithic France" sample, and a clade comprised of the former two samples is next closest to the "Etruscan series." The nearest outgroup is a clade that bifurcates into one subclade containing samples from Bronze Age England, Iron Age England, and Neolithic Denmark, and a second subclade that contains samples from Neolithic Russia and Neolithic Portugal.

As I recall, a sample of ancient Y-DNA from a Neolithic settlement in southern France has yielded Y-DNA from haplogroups G2 and I2a, both of which are also important in modern Sardinians. I do not know of any direct evidence to link ancient Etruscans with either G2 or I2a, but a hypothesis of such a link is plausible and should be investigated through analysis of Etruscan ancient Y-DNA.

eurologist said...

This mtDNA study confirms once again that the Etruscans were surely not derived substantially from Upper Paleolithic indigeneous hunter-gatherer populations of Europe, whose mtDNA was dominated by mtDNA haplogroup U4 and U5.


How can you come to this conclusion, based on the results of this paper? Also, how much support is there for your last statement re Southern Europe?

Grey said...

"Why were no other isolated Etruscan small towns sampled? Because the mtDNA there shows continuity."

Going by known later examples of intrusive Greek or Pheonician colonies amongst a native population wouldn't you expect a lot of admixture around the coastal colonies and mostly continuity further inland?

If the original Etruscans were an intrusive group that originally came by sea then i'd expect the same pattern with the collective Etruscans i.e. the invaders and the locals combined, after the coastal colonies take over the hinterland would be a mixed population of original Etruscans, mixed people and the original locals.

Basically wouldn't it be like trying to find the Pheonician signal in the region around Carthage - hardest at the points where it would originally have been the easiest?

Va_Highlander said...

@andrew: "(but not the Basque whose genetic distinctiveness and ethnogenesis probably have roots in a second wave of folk migration by Bell Beaker peoples in the Copper and early Bronze Ages)."

How so?

Crimson Guard said...

We already know that Italy was settled during the Neolithic(and earlier) and the population was effected very little during and after Roman times:






pconroy said...

Also remember that after the fall of Constantinople to the Ottoman Turks, many scholars and others fled to the city states of Italy, especially Florence, in the heart of Tuscany...

Simon_W said...

In the introduction the authors write that according to Livy the Etruscans were immigrants from north of the Alps. This however is a horrible mistake! Livy never made such a claim. This is just a modern interpretation, a questionable moreover. Livy just suggested that the Raeti were Etruscans who had fled from the Gauls. Usually this is interpreted in the sense that they had fled from northern Italy, from the Po Plain, whose colonization by the Etruscans and subsequent invasion by the Gauls is undisputed.

As for the evidence regarding Etruscan origins... There are several lines of evidence linking ancient Etruria, but also Umbria (!) with Central Europe. The Protovillanova culture, the relatively light pigmentation according to Biasutti's maps (based on Livi's data; Tuscany, also the formerly Etruscan area of northern Latium, and Umbria are in fact lighter than Emilia-Romagna and southern Latium + the formerly Faliscan territory), the predominance of R1b-U152, the fact that Tuscany clusters rather with northern Italy than with the south and (apparently) parts of central Italy... Now the Etruscan mt-DNA alines itself with this. And yet, as Dienekes mentioned, this does not exclude the possibility of an Anatolian origin of the Etruscan ethnos, as they could have assimilated and mixed with the locals.

Judging by the modern inhabitants, both the ancient Umbri and the Etruscans must have been very similar, which means at least one of them must have been formed by elite dominance.

Now the closest linguistic relatives of the Italic tribes were the Celts of central Europe, and the closest relative of Etruscan is Lemnian. I regard this as a hint.

Simon_W said...

Also, even if there may be some doubt about Herodotus' stories, Thukydides, who is generally considered the first serious historian, wrote that the Etruscans were Pelasgians, the same people who had inhabited Athens before the Greeks, and he added that only shortly before his birth, they had still been living in Chalkidiki and that even at his time, they lived in Lemnos - which is confirmed by the close relationship between Lemnian and Etruscan.

Simon_W said...

Sorry, a small lapsus: According to Thucydides they used to live on Lemnos and (at his time) still existed on Athos.

eurologist said...

"...and subsequent invasion by the Gauls is undisputed..."


The first settlements of cisalpine "Gauls" likely were not Gauls at all, but culturally Celtic people who probably spoke a language related to proto Celtic/Italic/Germanic - linguistically and ethnically (apart from the lack of admixture with "then-native Italians") not all that different from very early Italic speakers there.

True Gallic invasions came centuries later.

Simon_W said...

Well, I was referring to the invasions around and after 400 BC. There had been much earlier invasions, also by the Celtic Lepontii, among others.

The Etruscan presence in northern Italy can be divided into two stages: The Villanova culture around Bologna and Verucchio starting around 900 BC, and the later urban Etruscan colonization starting in the 6th century BC.

It's not entirely undisputed that the iron age Villanova culture already was Etruscan, but I tend to agree with this.

In any case, afaik the Etruscan Felsina was succeeded by the Celtic Bononia.

As for Livy's assertion: There is the objection that Raetic should be much more similar to Etruscan if it was true, i.e. if the Raeti had really seperated that late from the main body of Etruscans. But this doesn't prove that Livy's assertion has to be interpreted differently, i.e. as referring to a much earlier time, because it might be just a myth.