July 18, 2012

fastIBD over 2,257 Europeans

Razib points me towards a very interesting new paper that applies fastIBD over the large POPRES dataset of Europeans. The most interesting thing about this is that the authors develop techniques for estimating the time depth of the pattern of common ancestry across Europe, and hence are able to conclude that the Slavic expansion has played a bigger role in European history than the Germanic one.

A worthwhile improvement would be to apply a clustering algorithm like I did back in January over the fastIBD output; that way, one does not have to arbitrarily partition Europe into regions, but have the partitions jump out of the data.

A different idea to confirm the scenario presented in this paper would be to drill into different European populations. For example, in the case of the Italians, it would be worthwhile to identify whether there are particular sub-populations with likely Greek or Albanian ancestry who share an excess of IBD with modern Greeks and Albanians.

Population averages may mask such interesting patterns lurking in the data. For example, sub-clusters within populations can be identified with both fineSTRUCTURE and fastIBD, and the corresponding clusters can be assessed with supervised ADMIXTURE to detect how they differ from each other. For example, using this technique, I was able to infer 3 sub-clusters within the ethnic Greek population:

  • pop8 (mainland Greek) with ~23% North_European
  • pop11 (Greek Cypriot) with ~5% North_European
  • pop14 (Cretan, islander, mainland+Asia Minor) with ~12% North_European
  • I have also a strong hunch based on a few half Pontic Greek+half mainland Greek data points that unmixed Pontic Greeks would be related to pop22 (Northeastern Anatolia) with ~5% North_European
Based on these results and the fastIBD analysis of Ralph and Coop (the POPRES Greek sample is from northern Greece), it might appear that a hefty portion of the North_European component in Greeks may date to the medieval period, since it is relatively smaller in eastern Greeks and Cypriots and also in the South Italian/Sicilian cluster pop16 of a different analysis, with Italians as a whole lacking the eastern European affiliations of some Greek groups.

Interestingly, ~5% North_European levels would be similar to those of Armenians who are the closest linguistic cousins of the Greeks within the Indo-European family, as well as the the Anatolian Turkish cluster pop13 at ~9%.

Overall, it would appear that some mainland Greek groups received some input as the result of the medieval Slavic intrusions, since the mainland North_European excess appears as a "wedge" within the South Italy/Sicily/Crete/Anatolia/Armenia arc and the fastIBD pattern of sharing suggests that this is due to fairly recent connections.

As I have pointed out before, one limitation of the method of counting shared blocks of ancestry is that it does not disclose the directionality of gene flow. For example, gene flow between Germans and Slavs is detected in this study, which could be ascribed to Germans living in eastern Europe and/or to Slavs becoming acculturated Germans as a result of living within Germanic states or intermarrying with them prior to the age of the nation state.

Finally -and most interestingly- I hope that similar haplotype-based methods can be applied to a wider dataset, because, as it is becoming clear, Europe has not been isolated from Asia or Africa during its long history. The authors mention "Slavic or Hunnic" as an explanation for the pattern of shared ancestry in eastern Europe, but it is only by including Asian groups that we can detect the existence of real Hunnic (or Avar, or Mongol, or Pecheneg, or, ...) ancestry.

Moreover, I am confident that the Bronze Age is well within the power of haplotype-based methods to detect IBD. For example, South Asian populations clearly show differential patterns of affiliation with modern West Eurasian groups, most of which can date to no later than the Bronze Age. Together with the gradual incorporation of the new ancient DNA genomes that are bound to be coming our way soon, it seems that our picture of not only recent history, but also of late prehistory is bound to become much sharper.

arXiv:1207.3815v1 [q-bio.PE]


The geography of recent genetic ancestry across Europe

Peter Ralph, Graham Coop
(Submitted on 16 Jul 2012)

The recent genealogical history of human populations is a complex mosaic formed by individual migration, large-scale population movements, and other demographic events. Population genomics datasets can provide a window into this recent history, as rare traces of recent shared genetic ancestry are detectable due to long segments of shared genomic material. We make use of genomic data for 2,257 Europeans (the POPRES dataset) to conduct one of the first surveys of recent genealogical ancestry over the past three thousand years at a continental scale. We detected 1.9 million shared genomic segments, and used the lengths of these to infer the distribution of shared ancestors across time and geography. We find that a pair of modern Europeans living in neighboring populations share around 10-50 genetic common ancestors from the last 1500 years, and upwards of 500 genetic ancestors from the previous 1000 years. These numbers drop off exponentially with geographic distance, but since genetic ancestry is rare, individuals from opposite ends of Europe are still expected to share millions of common genealogical ancestors over the last 1000 years. There is substantial regional variation in the number of shared genetic ancestors: especially high numbers of common ancestors between many eastern populations likely date to the Slavic and/or Hunnic expansions, while much lower levels of common ancestry in the Italian and Iberian peninsulas may indicate weaker demographic effects of Germanic expansions into these areas and/or more stably structured populations. Recent shared ancestry in modern Europeans is ubiquitous, and clearly shows the impact of both small-scale migration and large historical events. Population genomic datasets have considerable power to uncover recent demographic history, and will allow a much fuller picture of the close genealogical kinship of individuals across the world.

Link

40 comments:

yisha3 said...

Do European Jews (Sephardim, Ashkenazim, etc) fare closer to pop 11 or pop 16 in such regards?

Or do they form their own subcluster?

aspromavro said...

Does that mean Slavic DNA is close to 20% in mainland Greeks??

Dienekes said...

We don't know exactly how North_European incoming Slavs would have been in Greece, nor do we know exactly how much North_European existed there before their arrival.

But, by observing populations from Sicily, Crete, Cyprus, Anatolia, and Armenia, and by knowing that the Slavic origin can be traced well to the north of Greece, it seems like a reasonable assumption that some of the North_European component arrived with them.

How much remains to be seen. My personal guess is like 10-20% for mainland Greece, and this is based on thinking that North_European was much lower (but not non-existent) in late antique Greeks, and that the Slavs would have been substantially North_European when they started to move, a little like their Baltic cousins who are like 90% North_European today.

Unknown said...

There's a lot more fun stuff to be done along these lines, as you say. Measurements of asymmetry in migration, and tying in a bunch more data (not to mention historical information) will be great.

A couple of comments: The partitions we used did actually jump out of the data -- plus some aribtrariness (needed because of fuzzy boundaries, plus low sample sizes like Slovakia), but a clustering algorithm gives somewhat different results depending on what data you choose to give it (e.g. what length blocks). It would be quite interesting to see what clusters one gets in various ways, and to figure out what they mean.

Pushing farther back in time (e.g. Bronze Age) runs up against the limits of IBD detection software like fastIBD -- there are a lot of segments, but one is a lot less certain about them, and the power to detect them will vary by population depending on allele frequencies. Finestructure demonstrates there's a lot of information at this scale, but we lack the models to be able to get good timing information out of it, so far.

cheers,
Peter Ralph

ps. It's Ralph and Coop (not Cook, oops =).

Dean said...

I think that 10-20% Slavic input in mainland Greeks is a reasonable estimate and agrees with Y DNA info.

The early Slavs seemed like easy converts to Christianity and Hellenism. I don't recall reading about Slavic resistance to Hellenism in Greece, such as that of Muslims rejecting conversion to Christianity or vice versa. This likely helped facilitate mixing with Greeks at a relatively early age. I read that some Slavs in Greece were angry at losing their independence to the Byzantines, but other than that and some initial, violent invasions, Slavs in Greece quietly and quickly converted to Hellenism and Christianity.

Dienekes said...

@Peter Ralph,

If you can release your .fibd.gz file, I can try clustering the individuals using the methods I've linked to.

eurologist said...

With regard to Germanic and Slavic expansion, one can make some interesting observations. Firstly, Germans share few ancestors within themselves, probably due to a large and not very homogenous population (Germans at the periphery tend to be rather similar to their neighboring populations). But this also means that if some group (say Eastern Germanic tribes) migrated elsewhere, they are not particularly closely related to Germans, today, and thus don't leave that type of signature behind.

Conversely, the Eastern Germanic tribes (and general population of that area) may have completely mixed in within the Slavic expansion, and even before one can use those terms: for example, the "BAL" component shares a huge number of ancient ancestors with Serbo-Croatians, but also with Germans. Basically, the region was thinly settled and that population does not exist in an identifiable way anymore, today, but had unusually large North-South diffusion and migration.

Nirjhar007 said...

Dear Dienekes, I am frankly telling you that if your suggestion on South Asia and Bronze age arrival of West Eurasians gives any practical evidence or direct evidence eg. aDNA then i shall quit the internet.
About components The age is vital not the proportion to fit it with any i mean any historic records not mere assumptions.
Have a good and truthful time.

Mr. Guy said...

@Dean

I'm no geneticist, but a more historically realistic estimate of Slavic admixture in Greeks is between 4-9%.

"The early Slavs seemed like easy converts to Christianity and Hellenism."

This is incorrect given that the Byzantine primaries mention a late conversion of Slavs (Cyril and Methodius in the 9th century). Furthermore, the process of "graikosas" that some Slavs underwent is not necessarily proof of widespread Hellenization of the Slavs (see the "Taktika" of Leo VI).

"I don't recall reading about Slavic resistance to Hellenism in Greece, such as that of Muslims rejecting conversion to Christianity or vice versa."

Although some Slavs were vassals of the Byzantine state and subject to some Christian Greek influences (Akameros of Thessaly), most Slavs were against any form of foreign rule (see Procopius). Furthermore, Slavic tributary status neither guarantees Slavic loyalty to the state nor substantiates any claims of an early entry into non-Slavic acculturation processes. Case in point, the Milingoi and Ezeritai were still Slavic when they rebelled against the Byzantine Empire (see "De Administrando Imperio"). Also, the notion that the Slavs were Hellenized early is based on a "logical absurdity" (as Douglas Dakin clearly states) where the Byzantine state had an army of teachers and priests to Hellenize and Christianize the Slavs (Dakin is correct as Byzantium's resources were limited prior to stabilizing its eastern frontiers from Persian and later Arabic invasions).

"This likely helped facilitate mixing with Greeks at a relatively early age."

Doubtful as the Slavs were a very minor demographic element in Greece and the Slavs were converted late. Furthermore, religious conversion does not always lead to cultural conversion (the Kievan Rus' and Bulgarians).

"I read that some Slavs in Greece were angry at losing their independence to the Byzantines, but other than that and some initial, violent invasions, Slavs in Greece quietly and quickly converted to Hellenism and Christianity."

The Byzantine primaries refer to the Slavs as raiders, bandits, and looters. Their invasions into Greece did not involve agriculturalists quietly settling the land. That the Slavs settled in the territories surrounding Thessaloniki is attested (Miracles of Saint Demetrios [with reservations]), but it is also attested that many Slavs were removed from the Balkans and sent to Asia Minor to fight against Byzantium's enemies in the East (i.e. Battle of Sebastopolis).

So, the genetic evidence may have a few kinks in its armor and Dienekes's estimate, though logical, is false. This is not to say that Greeks and Slavs did not mix, but that the mixing in question was quite negligible. If there was such a sudden or even gradual Greco-Slavic orgy, then it would have been attested in the Byzantine primaries (even the dubious "Epitomizer of Strabo" makes no mention of such an event) and there would have been clear evidence of Greco-Slavic physical hybridization (see Nikolas Xirotiris) and cultural hybridization such as a "Hellenized Slava".

And if you actually read the link, you'll find that even the authors (pp. 14-15) have their reservations: "Greece and Albania are also part of this putative signal of expansion, which could be because the Slavs settled in part of these areas (with unknown demographic effect)."


Mr. Guy

Dienekes said...

This is incorrect given that the Byzantine primaries mention a late conversion of Slavs (Cyril and Methodius in the 9th century).

Sts. Cyril and Methodius did not Christianize the Slavs of Greece, but rather were missionaries that went into Central Europe.

There are two types of Slav influences in Greece. One of them affected Macedonia and Thrace mostly, and consists of people who asserted a Greek ethnic identity even though they were of partial Slavic ancestry. Before the rise of nationalism in the Balkans, ethnic categories were more fluid, and it was no great event if a Slavic- or Albanian- or Greek-speaking Christian married someone of the other two communities. So, there are Macedonian Greeks who are descended from some ancestors who spoke Slavic (until quite recently), just as there are e.g., Bulgarians who are descended from ancestors who spoke Greek.

The other stream of Slav influence dates from the medieval period and is much broader, since the Slavs settled all the way to the Peloponnese. However, these were Hellenized through normal intercourse with their remaining Greek neighbors, and as a result of the imperial policy which resulted in resettlement of Greek and Armenian populations from Asia Minor in mainland Greece.

Dienekes said...

Addendum: Greeks from Sicily and Italy were also resettled in the Peloponnese once the region came back into imperial control.

pconroy said...

@Dienekes,

In terms of Sicily vs Greece and Balkans, my wife's grandmother has 23andMe relatives who are:
1. Greek
2. Serbian
3. Croatian

And one Italian with the name: Albano

Which could be related to Albania??

aspromavro said...

Dienekes, how do we know that the North_European excess is not a Mesolithic remnant closely connected with haplogroup I in the Balkans.

Mr. Guy said...

@Dienekes

"Sts. Cyril and Methodius did not Christianize the Slavs of Greece, but rather were missionaries that went into Central Europe."

True, but that still doesn't alter the fact that the Christianization of the Slavs by the Byzantine Greeks occurred late and not during, say, the late 6th and early 7th centuries when Emperor Maurice's brother, Peter, campaigned against the barbarians in the Balkans (see Theophylactus Simocatta).

"There are two types of Slav influences in Greece. One of them affected Macedonia and Thrace mostly, and consists of people who asserted a Greek ethnic identity even though they were of partial Slavic ancestry. Before the rise of nationalism in the Balkans, ethnic categories were more fluid, and it was no great event if a Slavic- or Albanian- or Greek-speaking Christian married someone of the other two communities. So, there are Macedonian Greeks who are descended from some ancestors who spoke Slavic (until quite recently), just as there are e.g., Bulgarians who are descended from ancestors who spoke Greek."

I respectfully disagree given that the various Christian communities (Greeks, Serbs, Bulgarians, etc.) mainly retained their distinct ethno-cultural identities during the Ottoman period as attested by multiple scholars including Thomas Harrison (see "Greeks and Barbarians"). Of course, I am not denying the existence of some admixture between these groups though in certain cases, the "admixture" may be conflated with "ethnolinguistic fluidity" in the sense that you had, for example, Albanian-speaking Greeks, Turkish-speaking Albanians, etc. And though language is an important element of ethnicity, it was not the sole determinant of ethnic identity for certain groups such as the Arvanites who have historically self-identified as Arvanitic-speaking Greeks or Greeks from Arbanon due to their possessing a pre-existing Byzantine Greek cultural orientation.

"The other stream of Slav influence dates from the medieval period and is much broader, since the Slavs settled all the way to the Peloponnese. However, these were Hellenized through normal intercourse with their remaining Greek neighbors, and as a result of the imperial policy which resulted in resettlement of Greek and Armenian populations from Asia Minor in mainland Greece."

This clearly echoes the "Chronicle of Monemvasia", which is neither a chronicle nor reliable (i.e. church propaganda). If anything, there is reliable primary evidence that directly contradicts the "flight of the Greeks" (Letters of Pope Gregory), as well as there being no evidence in the architectural record to substantiate the Chronicle's claim that Emperor Nikephoros I (r. 802–811) had the churches rebuilt.

As for the en masse resettlement of Sicilian and Asia Minor Greeks after the Slavic invasions, such an event(s) is not attested in the Byzantine primaries (except those sources that repeat, nearly verbatim, what is already stated in the Chronicle of Monemvasia). The author (or authors) of the Chronicle of Monemvasia used and misused a number of Byzantine primary sources (Procopius, Evagrius, Theophanes, Menander the Guardsman, and Theophylactus Simocatta) for the sake of furthering a church agenda involving the status of the metropolitan of Patras (the city of Patras, and not Monemvasia, was at the center of the Chronicle's "historical" narrative).

If anything can be taken as a certainty, it was that the geography of the Greek mainland was not conducive to any en masse Slavic settlement in contrast to the territories north of Greece (see Michael Whitby's "The Emperor Maurice and His Historian"). That a limited number of Slavs settled in parts of Thessaly and the territories surrounding Thessaloniki does not make all of Greece a "Slavonian land" as erroneously named by one Wilibald who never bothered to travel outside of Monemvasia in 732 to actually render his "Chronicle" somewhat reliable.


Mr. Guy

Dienekes said...

@Mr. Guy,

See p. 40 ff

http://www.arts.yorku.ca/hist/tgallant/documents/charanis.pdf

The genetic data is fairly clear about the existence of Slavonic influence in the modern mainland Greek population. The current study only confirms this by being able to date the admixture event, in agreement with the historical sources.

Albanian-speaking Greeks

There are no Albanian-speaking Greeks, or Slavic-speaking Greeks, or Vlach-speaking Greeks. Of course, it's understandable that some people are expansive in their use of the term "Greek", but I'm not one of them.

The fact that some Christian Albanians, Slavs, and Vlachs assimilated into the Greek nation and developed a Greek ethnic identity is of course obvious, and their affirmation of this identity has been mostly reciprocated by the Greeks. Today, through language education, social intercourse, and intermarriage, the old distinctions have all but disappeared. But, it serves no purpose to imagine that modern mainland Greeks all share the exact same genetic origin, and some of them just happened to speak a different language.

eurologist said...

how do we know that the North_European excess is not a Mesolithic remnant closely connected with haplogroup I in the Balkans

aspromavro,

Much of Northern Europe has significant diversity of various I subclades - the Northern Balkans not, at all. It is much more likely that these are remnant signatures from the East Germanic populations between the Oder and Vistula river, which are historically known to have expanded all the way to the Ukraine, and likely mixed with Slavic peoples and joined in the "Slavic" expansion. In particular, the I2a1b1a distribution conforms well to this expectation and a very recent founder effect.

Other I haplogroups are more specific to certain Eastern European regions and may be of older origin or may have connections Medieval Viking trade or German settlements there over the past 500 years (i2a2a?).

aspromavro said...

eurologist, what makes you think I2a1b1a was Germanic when it is not found at appreciable levels in any Germanic-speaking population?

Dienekes said...

The problem with I is that it's so darn limited to Europe. It's low even in Anatolia, and practically non-existent in most of the Caucasus. The Balkan type of I is also very low in Sicily and S Italy.

This opens to possible ideas:

1. It did exist in early Greeks but was limited to mountain Greek people from NW Greece that did not participate in city-state culture and did not found colonies.

2. It was absorbed by Greeks from their northern Balkan neighbors during antiquity or medieval period

Hopefully the BEAN and the Pinhasi project will throw some light onto the ancient Y chromosome gene pool in our part of the world.

Nirjhar007 said...

(I couldn't post the reply on the Iranian paper post after trying 100+ times!!! So please give your answer here or post the reply there and check the reply settings.)
Dienekes, Please give a link which gives data of J2s presence in Indo-arya uppercastes, as far as i know major brahmin groups don't have it which is a big hole to connect it with "I.E. Intrusion" and its denser in S India rather than N India.
About components again as i'm saying for eons the age is the real deal not the proportion! to connect them with any practical data or even theories.
Good times.

Dienekes said...

Not really my job to the literature survey for you. Your question is also off-topic. But, anyway:

"Haplogroup J2a-M410 is confined to upper caste Dravidian and Indo-European speakers, with little occurrence in the middle and lower castes."

http://dienekes.blogspot.com/2005/11/new-paper-on-indian-y-chromosome.html

"The frequency of this haplogroup is highest (23.5%) among the upper-ranked caste Brahmin and is lower (17.1%) among the middle-ranked caste Rajput. It is known that after the entry of the Aryan speakers into India, the Brahmins were the torchbearers and promoters of Aryan rituals (Karve 1961)."

http://dienekes.blogspot.com/2005/11/more-on-r1a1-age-and-haplogroup-j2-in.html

Kostas Oikonomou said...

Interesting. I wonder if there is a Slavic mark in the Greeks (or even Turks) of Bithynia, since Byzantines settled Slavs there.

So all the Greek mainland samples are from northern Greece? Do we know from which areas specifically?

eurologist said...

eurologist, what makes you think I2a1b1a was Germanic when it is not found at appreciable levels in any Germanic-speaking population?

aspromavro,

Because East Germanic tribes do not exist, any longer, and I2a1b1a is a very recently emerged, highly-derivative subgroup with demonstrably huge founder effects - similar to those (likely originally Ukrainian) Ria subgroups that characterize Slavic expansion at the 5-15% level (over those that are ancient in Central Europe).

Had I not labelled them "Eastern Germanic", but had simply said "long-time residents of the area between the Oder and Vistula rivers, and slightly beyond, during the fall of the Roman empire with excursions to Romania, sometimes farther south, often farther SE to the Ukraine, later joining Slavic expansion to the far East and into the northern Balkans" perhaps you wouldn't think twice about it.

But, as I alluded to further above, autosomal DNA exactly and completely confirms this. Why are Swedes, Balts, Germans, Poles, and Northern Balkans uniquely related through so many ancestors? Because they are tied to a climatically often very small, unique, but very homogenous population living in the center of that region, and climatically moving N/S over huge distances, as required. I am very confident a cluster analysis would exactly demonstrate what I am talking about.

Mr. Guy said...

@Dienekes

"See p. 40 ff"

I've read that particular work of Charanis a long time ago, but I respectfully recommend that you re-read the Byzantine primaries and also consult with more recent scholarship such as the "The Emperor Maurice and His Historian" by Michael Whitby, "The Making of the Slavs" by Florin Curta, "A History of Byzantium" by Timothy Gregory, "Oxford Dictionary of Byzantium" by Alexander Kazhdan, etc.

Also, you may want to do some research on what exactly the "Chronicle of Monemvasia" is rather than believe everything Peter Charanis says about it ("absolutely trustworthy"). It's pretty obvious that Charanis relies on the "historicity" of the "Chronicle" to justify his claims even though the Byzantine primaries that the "Chronicle" relies on make no mention of Slavs settling in Greece on a massive scale. Even the toponyms "discovered" by Max Vasmer appear to be dubious given that they are surprisingly absent from the Byzantine literary (except for "Avarino" and "Sclavohori"). And even if these toponyms existed, they don't necessarily serve as direct evidence of Slavic settlement (as clearly stated by Vacalopoulos's "Origins of the Greek Nation"). So forgive me if I'm not really convinced by your assertion that the IBD study somehow proves your personal estimate of 10-20% admixture when the authors of the IBD study have explicitly mentioned their reservations ("unknown demographic effects") and have explicitly mentioned the fact that the Slavic genetic impact on Greeks was quite limited.

"The genetic data is fairly clear about the existence of Slavonic influence in the modern mainland Greek population. The current study only confirms this by being able to date the admixture event, in agreement with the historical sources."

I mean no offense, but you appear to operate under the false assumption that I'm denying the existence of admixture between Greeks and Slavs just because I disagree with your personal estimate of 10%-20% admixture. But even if the "genetic data" more or less concurs with your personal estimate, that doesn't necessarily mean that the data in question is entirely accurate ("false positives"), devoid of limitations, or that it concurs 100% with other forms of evidence (the authors of the IBD study talk about false positives and the need for corroborative evidence).

All the IBD study confirms, as far as I know, is that the Slavs had a negligible genetic influence on both Greek and Albanian populations (pp. 14-15), which coincides with the significant dearth of Slavic cultural elements in Greece as attested by multiple scholars including Alexander Kazhdan. And though you are entitled to publish any estimates you see fit (it's your blog after all), that doesn't mean that you're numbers are always reflective of historical actuality.

"There are no Albanian-speaking Greeks, or Slavic-speaking Greeks, or Vlach-speaking Greeks. Of course, it's understandable that some people are expansive in their use of the term "Greek", but I'm not one of them."

As surely as there were/are no English-speaking Greeks, or Swedish-speaking Greeks, or Turkish-speaking Greeks, etc. Even your country's first prime-minister, Kapodistrias, was born in Venetian-held Corfu and was in the Russian foreign service during the reign of Alexander I (r. 1801-1825). Was he not an Italian-speaking and Russian-speaking Greek?

Cheers and thanks for the chat.


Mr. Guy

P.S. By the way, why are your articles on Greek DNA unavailable?

aspromavro said...

Why are Swedes, Balts, Germans, Poles, and Northern Balkans uniquely related through so many ancestors?

That BAL and SC result you mentioned strikes me as very strange and dubious. Maybe Dienekes could try to replicate it on the Dodecad set?

Dienekes said...

I've read that particular work of Charanis a long time ago, but I respectfully recommend that you re-read the Byzantine primaries and also consult with more recent scholarship such as the "The Emperor Maurice and His Historian" by Michael Whitby, "The Making of the Slavs" by Florin Curta, "A History of Byzantium" by Timothy Gregory, "Oxford Dictionary of Byzantium" by Alexander Kazhdan, etc.

It does not particularly matter if scholarship is recent or not. Florin Curta for example, imagines that the Slavs were "made" by the Byzantines, and that the Slavic tongue spread as a lingua franca of the Avar khaganate. He proposes that the "spread of the Slavs" is illusory, and the spread of the "Slavic language" was a cultural phenomenon. The discovery that Eastern Europeans consistently share an excess IBD tracts dating to medieval period disproves that hypothesis.

So forgive me if I'm not really convinced by your assertion that the IBD study somehow proves your personal estimate of 10-20% admixture when the authors of the IBD study have explicitly mentioned their reservations ("unknown demographic effects") and have explicitly mentioned the fact that the Slavic genetic impact on Greeks was quite limited.

I did not assert that _the study_ has that estimate. That is my personal estimate based on the excess of North_European in the mainland Greek population relative to non-mainland Greeks, S Italians/Sicilians, Anatolians of various kinds, that clearly points towards a local excess of such ancestry. Depending on one's assumptions about how much North_European there was in Greece pre-6th century, and how much North_European there was in early Slavs, one can come up with different numbers, and to repeat what I actually said:

"How much remains to be seen. My personal guess is like 10-20% for mainland Greece, and this is based on thinking that North_European was much lower (but not non-existent) in late antique Greeks, and that the Slavs would have been substantially North_European when they started to move, a little like their Baltic cousins who are like 90% North_European today."

All the IBD study confirms, as far as I know, is that the Slavs had a negligible genetic influence on both Greek and Albanian populations (pp. 14-15)

The study in no way characterizes the amount of genetic influence as "negligible".

P.S. By the way, why are your articles on Greek DNA unavailable?

I no longer maintain those. You can replace 110mb or 50webs with ifreepages or awardspace, there ought to be mirrors over there.

sds said...

@ eurologist said: "With regard to Germanic and Slavic expansion, one can make some interesting observations. Firstly, Germans share few ancestors within themselves, probably due to a large and not very homogenous population (Germans at the periphery tend to be rather similar to their neighboring populations). But this also means that if some group (say Eastern Germanic tribes) migrated elsewhere, they are not particularly closely related to Germans, today, and thus don't leave that type of signature behind.

Conversely, the Eastern Germanic tribes (and general population of that area) may have completely mixed in within the Slavic expansion, and even before one can use those terms: for example, the "BAL" component shares a huge number of ancient ancestors with Serbo-Croatians, but also with Germans." Is it possible that this is the case because many so-called Germanic tribes, especially those labeled East Germanic, were in reality Slavic (or proto-Slavic) in totality, or at least in some some substantial proportion?

rohamius said...

Hi,
What about the arabic/semitic influences/admixtures in Crete, Sicily, Cyprus or Pontus? Why these places should be considered as "clean Greek" per genetic samples? What I mean is that the percentage influx of North Europeans in Mainland Greece might not be totally of Slav invaders as the percentage of North Europeans in Crete, Sicily, Cyprus etc might have been altered by influx of other races in these regions that would bring down the percentage of North Europeans.

Dienekes said...

What about the arabic/semitic influences/admixtures in Crete, Sicily, Cyprus or Pontus? Why these places should be considered as "clean Greek" per genetic samples?

This can also be quantified:

http://dodecad.blogspot.com/2012/02/correspondence-between-chromopainter.html

For example, pop8 has 9% SW_Asian and pop13 has 11%, pop22 has 9%, even Cypriot pop11 18%.

Overall, it seems that the SW_Asian component which is modal in Arabs and well-represented in Semitic populations does not vary a great deal in these groups. This is also in agreement with Y chromosomes where there is a limited contribution of such lineages.

More than a 4-fold reduction in the levels of North_European would require substantial levels of admixture. This would have to happen simultaneously in a broad arc of "southern populations" from Italy to Armenia. Also, we have the evidence about the date of this admixture. Overall, on balance, an increase in the N European from a non-zero base seems like a more parsimonious explanation.

eurologist said...

Is it possible that this is the case because many so-called Germanic tribes, especially those labeled East Germanic, were in reality Slavic (or proto-Slavic) in totality, or at least in some some substantial proportion?

sds,

Archeology, historic references, linguistics, and substantiating genetic results tell us otherwise.

For one, archeological continuity exists and evidently united northern German areas with those of at least Western Poland east of the Oder, if not (by most scholars) the Vistula on both shores - at least 2,000 - 4,000 years ago.

Historic references name eastern tribes east of the Oder to the Vistula and all the way to to Romania and the Ukraine ~2,500-1,500 years ago as clearly Germanic.

There are very ancient Proto-Germanic attestations into Uralic, much earlier than into or from Slavic. Early Germanic tribes clearly bordered Uralic ones, and surviving German-Slavic language exchange - while plenty - can be traced to be exclusively rather recent.

Finally, both autosomal and y-DNA genetic results have been confirming this picture from day one.

Nirjhar007 said...

http://www.pnas.org/content/103/4/843/F2.large.jpg
Similarly with Thangaraj et al.(2010) The distribution of y-dna J2 is weak on the Eastern half of the whole country now the Sahoo et al. 2005 also provides as you can see the concentration of Hgs but the J2 is only concentrated on the West Southern India and as Mukherjee et al. Have noted its worthy presence in S Indian Brahmins, but incase of West Bengal Brahmins, Bhumihar Brahmins (Bihar), Uttarpradesh Brahmins, Utkala Brahmins, Karnataka Brahmins etc. the Presence of J2 is 0 but R1a1,R2a and H are present!! Check the Wiki.
The Genealogical rate date of J2a is ~800b.c In 800b.c. There was an archaeologic intrusion in The Indus Basin and a part of Chitpavan etc. South western Brahmins are noted to be of Recent Indo-scythian/Iranian origin so the J2a is clearly young than the AMT plot in India.
Good times.
P.S. What if Farmana DNA provides J2a with R1a1, R2a etc? The story will be quite different with direct evidence.
Ta ta... N.

Dienekes said...

@Nirjhar007

1) You are OFF TOPIC.
2) Posting a synthetic frequency map is not the same as posting actual data. You claimed, for example, that Bihar Brahmins don't have J2 and they do (Sahoo et al. 2006). In fact all of the groups labelled Brahmin in that large study possess J2.

Posting erroneous information is a big no-no if you want to continue commenting on this blog. I don't doublecheck most of the assertions posted on this blog, but if you want to contradict me, you'd better bring it.

The Genealogical rate date of J2a is ~800b.c In 800b.c.

No, it is not:

http://dienekes.blogspot.com/2008/08/origin-of-hindu-brahmins.html

Nirjhar007 said...

Sorry for the wrong info, i will talk on that link.
Good times.

Unknown said...

There's something dubious on your conclusions here Dienekes.

First of all, in analyzing ancient historical migrations it is very logical that we should expect more North European ancestry in mainland Greeks.

We know for example that the proto-hellenes were centered in Epirus, Thessaly and Macedonia. From there they spread througout what came to be known as the Greek world. During that time the Pelopponesse, Crete and the Aegean were more densely inhabited by pre-hellenic peoples who probably had less North-European ancestry.

The Greeks who colonised Italy were mostly from the South(Pelopponesse) and had to this day have less North European ancestry than other mainland Greeks. They mixed with the native Italians who in turn also mixed with Phoenicians, so the percentage of North European haplogroups declined.

Same can be said for Asia minor Greeks and Pontians.

My point is that there was genetic diversity in Ancient Greece. Central and Northern Greece were more Northern European, while Southern Greeks (the bulk of the Ancient Greek colonisers)were mostly descendants of Neolithic pre-Greeks.

Dienekes said...

We know for example that the proto-hellenes were centered in Epirus, Thessaly and Macedonia. From there they spread througout what came to be known as the Greek world. During that time the Pelopponesse, Crete and the Aegean were more densely inhabited by pre-hellenic peoples who probably had less North-European ancestry.

You can walk from the Peloponnese to Thessaly in a few days. It's not like these regions are in completely different latitudes, they're very closeby in the wider European context.

My point is that there was genetic diversity in Ancient Greece. Central and Northern Greece were more Northern European, while Southern Greeks (the bulk of the Ancient Greek colonisers)were mostly descendants of Neolithic pre-Greeks.

That's speculation in the absence of data. Actually, the only data we have is the absence of any mention of phenotypic differences between Greek peoples; if some types of Greeks were much more North European than others, then this would have been noticed and commented on by the ancient writers, but nothing like that has survived in the ancient literature.

Unknown said...


@Dienekes

It's not entirely speculative. In fact, based on the historical demograohics of Ancient Greece, I would say it is speculative to assume that all Greeks looked exactly the same. I.e. that the Greeks of Cyprus looked exactly the same as the Greeks of Macedonia or Thessaly.

As you pointed out, you can walk from the Peloponesse to Thessaly in a few days. Well, within a few days you can walk from Thessaly to Thrace. And we do have accounts that Thracians looked more northern compared to Greeks. Since all paleo-balkanic people lived in close geographical proximity and we have accounts that Thracians looked different from Greeks, it is therefore logical to assume that there may have been gradual differences in the complexion of Greeks and that the mainland Greeks of Thessaly, Boaetia, Epirus, Macedonia were something in between Southern Greeks and Thracians. Ofcourse the differences between Northern and Southern Greeks, like today, were not extreme and not considerable enough to mention by the Ancient writers.

Having said that, I don't deny that medieval migrations may have had an impact as well. But we shouldn't underestimate some genetic diversity in Ancient Greece as well.

ps said...

Pontic Greeks and Cypriot Greeks both have only 5% north european.

Those two populations had no links with Greece's northern neighbours during antiquity nor during slavic invasions.

Concerning Greeks (today Greeks, ancient Greeks and medieval Greecs), it seems very important to have results from all regions separately (Creta, Asia Minor, Thrace, Cyprus, Rhodes, Pontic Greeks etc) because it is obvious that the results are not the same. It is very interesting to see that those populations have a very strong common identity based on common language, religion, history despite of invasions and conquests. This shows the strenght of Greek culture.

A little bit of genetic diversity is not a problem when identity is strong - with ancient Greek roots of course, because even if there have been some mixtures, there is no Greek identity, language, culture, if there is no ancient Greece.

This identity already existed in the region before slavic or Turkish identity. I wonder if Greek identity will manage to be as strong as this with the new immigrants coming in Greece from islamic countries (I know that many muslim pomaks feel they are Greeks but I don't think pomak culture cannot be compared with Afghan and Pakistanese culture can it?).

Concerning slavic invasions, most searchers now admit that East Roman Empire (sorry I don't like the terms "Byzantine empire" which means nothing) deported some slavic tribes from Greece to Asia Minor and some Christians from Asia minor (Greeks, Armenians) to Greece. Is there any genetic evidence of this?

Concerning Cyprus, there are so many Russians in Cyprus now... This is the new slavic (peaceful) "invasion" which will change Cyprus genetics in a few years : )

Constantine said...

>> Pontic Greeks and Cypriot Greeks both have only 5% north european.

Cyprus was colonized by Phoenicians as well so it is logical that even in ancient times there must have been some differences between mainland Greeks and Ancient Greeks. Cypriots having less Northern European Ancestry.

Also, Cyprus was never invaded by the Dorians. They could have brought more Northern European ancestry with them.

In Crete you had the Minoans who had more Near Eastern Admixture compared to mainland Greeks too. This may also be true for the islanders.

I would say that Central and Northern Greece had more Northern European admixture than other parts of the Greek world, and that some more came during the middle ages.

We will have to wait and see some more results on the rate of diversity of Ancient Greeks. In any case history surely suggests there may have been some.

ps said...

Answer To Constantine :

No, there were not many Pahenicians in Cyprus. Juste some Phoenician counters. They never colonized Cyprus. Juste some little counters.

The most important population in Cyprus were Eteo-Cypriots, not Phaenicians. Most of them were from Asia minor. Then they were assimilated, absorbed by the Achaeans

In Crete there were the Minoans for sure. Assimilated, absorbed by populations from Mainland Greece.

And what about Greeks from Asia minor? Phrygians, Carians. Assimilated, absorbed by populations from mainland Greece.

In mainland Greece, Thracians, Assimilated, absorbed by Greek populations.

There is always a common point : some Greek populations came to a non-Greek Areas and those Greeks assimilated those populations. They absorbed them. And the result of this was a greater Greek population.

Let's say all those populations have a father from the Greek family, a mother from a non Greek family (or the contrary) and then the whole family becomes Greek.

Also, there were no Slavic, nor illyrian admixture in Cyprus.

But don't forget that when Eastern Roman Emperors fought against the Slavic populations that came in Greece, they deported some of those sLAVES in Asia Minor, and they deported some Asia Minor populations in Greece.

So Greek populations of Greece also come from Asia Minor.

Anaxagoras said...

This is an extremely interesting post. I am particularly interested about the genetic admixture of Greek Cypriots and their differences/similarities with mainland Greeks. I totally agree with all the comments highlighting the 'Northern European' component as the one distinguishing the different ethnic Greek population subgroups. I would also add to that the 'East European' component, which (as with the North Eur component) is highly associated with the Slavic expansion. If you look closely at the Dodecad k12b raw data, you will notice that the main difference between Greek Cypriots and mainland Greeks is the huge difference in the 'North European' component (4.5% vs. 20.2%, respectively). Similarly, when looking at the Dodecad v3 raw data, mainland Greeks have an 'East European' admixture of 11%, while for Greek Cypriots it is just 3.4%. Another major difference is the Cucausus component which is much more dominating among Greek Cypriots (49.3%) compared to mainland Greeks (37.4%). Just a final comment, there is a very bizarre lack of western European admixture among the dv3 Dodecad Cypriot sample (1.3%!). I will write a separate post about this issue however.

Anaxagoras said...

This is an extremely interesting post. I am particularly interested about the genetic admixture of Greek Cypriots and their differences/similarities with mainland Greeks. I totally agree with all the comments highlighting the 'Northern European' component as the one distinguishing the different ethnic Greek population subgroups. I would also add to that the 'East European' component, which (as with the North Eur component) is highly associated with the Slavic expansion. If you look closely at the Dodecad k12b raw data, you will notice that the main difference between Greek Cypriots and mainland Greeks is the huge difference in the 'North European' component (4.5% vs. 20.2%, respectively). Similarly, when looking at the Dodecad v3 raw data, mainland Greeks have an 'East European' admixture of 11%, while for Greek Cypriots it is just 3.4%. Another major difference is the Cucausus component which is much more dominating among Greek Cypriots (49.3%) compared to mainland Greeks (37.4%). Just a final comment, there is a very bizarre lack of western European admixture among the Dodecad v3 Cypriot sample (1.3%!). I will write a separate post about this issue however.