search this blog

Wednesday, June 20, 2012

First direct evidence of genetic continuity in West and Central Poland from the Iron Age to the present

I've just been sent a fascinating thesis on the mtDNA of Iron Age and Medieval samples from Poland. It suggests direct genetic continuity between Iron Age samples belonging to the Przeworsk and Wielbark Cultures, of what is now West and Central Poland, and present-day Poles. Here's the English summary, and a map of the sites under study:

For many years the origin of the Slavs has been the subject-matter in archaeology, anthropology, history, linguistics and recently also modern human population genetics. By now there is no unambiguous answer to a question where, when and in what way the Slavs originated. For the purposes of this dissertation, the analysis of ancient human mitochondrial DNA was applied. The ancient DNA was isolated from 72 specimens which came from Iron-Age and medieval graveyards from the area of current Poland. Ancient mtDNA was extracted from two teeth from each individual and reproducible sequence results were obtained for 20 medieval and 23 Iron-Age specimens. On the basis of HVR I mtDNA mutation motifs and coding region SNPs each specimen was assigned to a mitochondrial haplogroup. The obtained results were used together with other ancient and modern populations to analyse shared haplotypes and population genetic distances illustrated by multidimentional scaling plots (MDS). The differences on genetic level and quite high genetic distances (FST) between medieval and Iron-Age populations as well as significant number of shared informative haplotypes with Belarus, Ukraine and Bulgaria may evidence genetic discontinuity between medieval and Iron Ages. From the other side, the highest number of shared informative haplotypes between Iron-Age and extant Polish population as well as the presence of subhaplogroup N1a1a2, can confirm that some genetic lines show continuity at least from Iron Age or even Neolithic in the areas of present day Poland. The results obtained in this work are considered to be the first ancient contribution in genetic history of the Slavs.

Below is an MDS from the thesis, based on data corrected for the effects of potential relatives in the Iron Age sample. I don't think it's a particularly useful way of judging the intra-European affinity of the two ancient Polish groups, mostly because the samples are small, and contemporary North, Central and East Europeans don't differ very much in terms of mtDNA. Nevertheless, we can see that both the Iron Age (Okres Rzymski) and Medieval (Sredniowiecze) samples fall within the range of modern European mtDNA diversity. On the other hand, the German Neolithic LBK sample (Neolit LBK Niemcy) clearly does not, because it's sitting at the far right of the plot, away from the main European cluster. This dichotomy between the genetic structure of the LBK farmers and modern Europeans has been demonstrated in previous studies, but the reasons for it are still a mystery.

Interestingly, modern Poles are closer to an Iron Age sample from Denmark (Okres Zelaza Dania) than to the Polish Iron Age set. However, as per the summary above, the author also compared the frequencies of the most informative haplotypes among the modern and ancient samples, and found that extant Poles are the closest group to the Polish Iron Age remains, followed by Balts, Swedes and Baltic Finns. Below is a table showing those results.

According to the author, these matches might hint at Baltic, Germanic and Finno-Ugric influences in the Polish Iron Age population. Perhaps, but in my opinion, they're simply in line with geography, and reflect the general North European character of maternal lineages shared by populations from around the Baltic, both today and during the Iron Age.

The results for the Medieval Polish sample are more intriguing, because they're somewhat out of whack with geography. Its best matching modern groups are Belorussians, Ukrainians and Bulgarians. This might suggest that, during the early middle ages, the territory of present day Poland experienced an influx of groups from what are now Belarus and Ukraine, who then melted into the gene pool of the natives of Polish Iron Age descent. However, conversely, it might mean that Belorussians, Ukrainians and Bulgarians descend in large part from fairly specific medieval groups from the area of modern Poland.

In any case, whether present day Polish territory saw some migrations from the immediate east during the Medieval period or not, this preliminary look at ancient Polish mtDNA suggests long-standing genetic continuity in the region. What it clearly doesn't show is a complete, or almost complete, population replacement in the areas between the Oder and Bug rivers during the migration period.

Indeed, the thesis results put into doubt past notions that the Przeworsk and Wielbark cultures were of Germanic origin.

The (mtDNA) haplogroup missing from both the Iron Age and medieval samples from the territory of modern Poland was haplogroup I. In contemporary Slavic populations, this haplogroup is found at levels ranging from 1.2% in Bulgarians to 4.8% in Slovaks. It was also recorded at high levels in ancient remains from Denmark. It showed a frequency of 12.5% in an Iron Age sample, and 13.8% in a medieval sample. Melchior et al. 2008 suggest that haplogroup I might have been more common in Denmark and Northern Europe during that period. Therefore, the lack of this haplogroup in ancient DNA from the territory of modern Poland, might mean that the Przeworsk and Wielbark cultures should not be identified with Germanic populations.

I'm sure more ancient DNA studies are on the way looking at the origins of Slavs and Poles. Indeed, if the Y-chromosomes of Przeworsk and Wielbark remains are successfully tested, I won't be surprised if they look fairly typical of modern Poles, with a decent representation of R1a1a-M458, which is the most common Y-chromosome haplogroup in Poland today.

Anna Juras, Etnogeneza Słowian w świetle badań kopalnego DNA, Praca doktorska wykonana w Zakładzie Biologii Ewolucyjnej Człowieka Instytutu Antropologii UAM w Poznaniu pod kierunkiem Prof. dr hab. Janusza Piontka


Michał said...

I would like to add a comment to your comment, cited below:

"This might suggest that, during the early middle ages, the territory of present day Poland experienced an influx of groups from what are now Belarus and Ukraine, who then melted into the gene pool of the natives of Polish Iron Age descent. However, conversely, it might mean that Belorussians, Ukrainians and Bulgarians descend in large part from fairly specific medieval groups from the area of modern Poland."

The second option seems to be highly unlikely, as it would require a (relatively) massive post-medieval (or late medieval) migration from Poland towards east (Belarus, Ukraine) and south (Bulgaria, and generally Balkans) that was characterized by extreme selectivity towards specific mtDNA haplogroups. No such late “selective” migration is documented in historical records (especially as regards the hypothetical movement from Poland to the Balkan peninsula). Alternatively, one would have to assume a selective extermination of those "Eastern " mtDNAs in a post-medieval Poland, which seems even less likely. And after all, none of those scenarios explains how these “Eastern” mtDNAs arrived to the medieval Great Poland in the first place, as they are supposed to be absent (or much less frequent) there during the earlier Przeworsk period.

Thus, we are left with two possibilities only. Firstly, there could have been a significant early medieval mtDNA influx from East and/or South that was later followed by another (late medieval, modern?) influx of Przeworsk-like mtDNAs, either from outside (Germany?) or from some internal enclaves (Southern Poland, Pomerania?) that were not tested in this study. Secondly (and most likely in my opinion), the differences recorded in this study are not statistically significant, so they cannot be used to prove or disprove any theory regarding the Slavic ethnogenesis.

Davidski said...

^ Like you say, the data doesn't prove or disprove anything about Slavic origins or direction of medieval migrations, due to the small samples. So Poland might indeed be the source of Slavic migrations both to the east and south.

But what I found really interesting was the lack of any clear links between the Iron Age Poles and modern Germans, and more importantly, between the Iron Age Poles and ancient Danish samples (for example, no haplogroup I in the ancient Polish mtDNA).

Based on these results, it's really difficult to see Przeworsk as a Germanic culture.

Michał said...

As for the relatively high frequencies of haplogroup I in the Iron Age and Medieval Denmark, these values were estimated based on equally small samples. Most importantly, the frequency of this haplogroup among modern Danes is at the same level as in modern Poland (or even slightly lower), which indicates that those aDNA data were most likely not representative for the entire country (not to mention all Germanic tribes), as there were not many post-medieval waves of non-Germanic immigrants coming to Denmark and potentially contributing to such extensive dilution of this putative "Germanic" haplogroup.

It seems to me that your interpretation is not balanced enough, as you seem to be focused on searching for any data that are consistent (or actually not inconsistent) with your view on the origin of Slavs, while at the same time you seem to neglect all data that pose some problems to that theory (like the suggested early medieval influx of mtDNA from the East)

Also, when you write that "what I found really interesting was the lack of any clear links between the Iron Age Poles and modern Germans", it is hard to imagine how those "special" links should look like, especially when we know that the difference between modern Poles and modern Germans is not statistically significant (even when tested on large samples!).

Finally, I find it very strange when you admit that the sample is small and the observed differences are not significant, yet you state that "Based on these results, it's really difficult to see Przeworsk as a Germanic culture". In fact, when you take a look at Fig. 14 in that paper, you will see that the Przeworsk sample is located much closer to the German and Swedish samples than to any Slavic sample (including Polish). Excluding some potentially related individuals from the Przeworsk sample lead to the slightly different outcome (Fig. 15), but even then the Przeworsk sample was located closer to modern Swedes than to modern Poles. I am fully aware that all these differences are not significant, but this is just an example of how an equally biased interpretation (though in another direction) may look like.

Please don't get me wrong. I really appreciate your fabulous work on autosomal DNA, but you seem to lose your objectivity as soon as the question of Poland and the origin of Slavs arises.

Davidski said...


"Special" links between the Iron Age Poles and modern Germans would most likely be seen here as informative haplotype matches between these two groups. But there are no such matches (the table I attached shows zero).

As for the MDS, if you look at it closely, it seems that the Iron Age and medieval Poles are nearest to Balkan Slavs. I think the reasons for this are twofold. Firstly, using MDS on broad mtDNA haplogroup frequencies isn't a good way to discriminate at fine scale between closely related North/Central/East European groups. Secondly, it seems that dimension 1 deals with similarity to the German Neolithic LBK sample, and dimension 2 is influenced heavily by diversity, with modern Finns probably showing the lowest diversity, and modern Germans the highest.

So this would suggest that Polish mtDNA was more varied in the past, but has since lost some of that variance, most likely due to several major population crashes, like during the Great Northern Wars.

Also, I really do think the lack of haplogroup I is important. The samples from ancient Denmark weren't that small, and they produced basically the same results from two burials separated by hundreds of years. I think that if the Iron Age Poles had a close genetic relationship with Iron Age and medieval Danes, then they'd show at least some haplogroup I.

Here's what the authors of that paper on the ancient Danish mtDNA theorized...

"The overall occurrence of haplogroups did not deviate from extant Scandinavians, however, haplogroup I was significantly more frequent among the ancient Danes (average 13%) than among extant Danes and Scandinavians (~2.5%) as well as among other ancient population samples reported. Haplogroup I could therefore have been an ancient Southern Scandinavian type “diluted” by later immigration events."

Let's wait for more results, including ancient Y-DNA. Suffice to say, I'm not expecting that I'll be saying anything different to what I said in this blog post when these results come out.

Michał said...

I can then conclude from your answer that the fact that none of the 11 “informative” Przeworsk haplotypes was found among 300 modern Germans clearly suggests that Przeworsk could not have been Germanic (though Przeworsk has never been considered as ancestral to modern Germans, even by the strong proponents of its Germanic status, as Germanic does not mean German), while the similar fact that none of those 11 haplotypes was found in 20 Medieval (Polish) samples from the very same region (Great Poland) does not prevent you from writing a post titled “First evidence of genetic continuity in Poland since at least Iron Age”. Do you really feel that this is a fair interpretation of all this data?

Let’s go back to your theory that Przeworsk (or Pre-Medieval Poland) was a source of the Slavic expansion towards east and south, although the mtDNA data clearly suggest that it was the other way around (i.e. Poland was rather a target of the mtDNA influx from the East and/or South). How would you explain the fact that the Medieval mtDNA sample form Poland shows clear association with modern Belarus, Ukraine and Bulgaria, while no such association is found for the Iron Age (i.e. Przeworsk) sample? In fact, the Przeworsk sample shows statistically significant difference when compared with modern Ukraine and Belarus, while no such difference is observed when comparing Przeworsk with modern Germany and Sweden, or with the Iron Age and Medieval Denmark. The only sensible interpretation of this result is that Przeworsk had nothing to do with the Slavic expansion, while Early Medieval Poland (or at least Great Poland) was an obvious target of this expansion.

When discussing the importance of haplogroup I as a putative Germanic “marker”, you avoid to answer my major objection, mainly that the frequency of haplogroup I in modern Denmark, Sweden and Germany does not exceed the frequency of this haplogroup in Poland, while it reaches its peak among Carpathian Lemkos and on the Krk island in Croatia (11% in both cases). How would you explain the fact that this “Germanic marker” shows no correlation with Germanic-speaking populations.

The explanation provided by the citation from the original paper (Melchior et al.) is not satisfactory, as there are no records of any post-medieval immigrants who would be responsible for the nation-wide decrease of the haplogroup frequency by the factor of six. Actually, I have just read this paper, which only confirmed my initial suspicions that the ancient sample was not representative for the whole country. All these Iron Age and Medieval samples were taken from the relatively small region of Denmark (two islands only, Zealand and Funen), with six haplogroup I samples found on Zealand and one on the nearby island of Funen. There were no samples from Jutland, which constitutes the major part of the country. Importantly, it is the Northern Jutland region that shows some clear association with Przeworsk according to many archeologists.

Davidski said...


Yes, I think the lack of informative haplotype matches between the Iron Age Poles and modern Germans is very important. It's also important that the Iron Age Poles show the highest number of informative haplotype matches with modern Poles.

So at the moment, I believe these Iron Age Poles played a part in the ethnogenesis of West Slavs in what is now Poland.

The medieval mtDNA results make me think that there was an intrusive element from the immediate east in Poland at that time, living alongside the direct descendants of the Polish Iron Age. These two groups probably eventually mixed, creating the modern Polish gene pool. Although the problem with that scenario, is that there's a lack of the East Slavic R1a clade in Poland. So ancient Y-DNA from Poland is sorely needed at this point.

As for the other stuff, like the haplogroup I, the scientists thought it was worth mentioning, so I mentioned it as well. Obviously, these people aren't total morons, so when they see something of note, we should consider it as well.

wilmar said...


a.) Wielbark n=19 12xH (5 similar/identical), 3xW, 2xU, 1x J,T
Przeworsk n=4 2xH, 1xU,N

b.) Polish sample n=413-113 = 300 ("Jeśli liczba osobników z populacji łączonych była znacznie większa od 300, wówczas w sposób losowy wybierano pulę 300 osobników. Wybór losowy przeprowadzano w przy użyciu formuły „random factor” w Excel 2010.")
Pomerania-Gdańsk n=166
Upper Silesia n=87
Kaszuby n=87
Suwalszczyzna n=73

"I won't be surprised if it looks fairly typical of modern Poles, and includes R1a1a-M458, which is the most common Y-chromosome haplogroup in Poland today."

And R1a-Z280.:)

Unknown said...

Iron Age Denmark closer to Poland by DNA than Iron age Poland is to Poland!?

Is this perhaps evidence of the Lechitic Slavic Polabians being the inhabitants of Denmark into the Iron age!?

As I am sure you know.... Polabians were Lechitic Slavs & thus like a Western Polish tribe in the area of North-Germany - Denmark border region.

But... I mean this DNA evidence seems to in my eyes support that Polabians were inhabitants of Denmark during the Iron age!

......... As for the Iron age map of Germany.

It looks like Poland & Germany are closer in the Modern era than in the Iron age!

This makes sense though giving how much during the Medieval era that Germans have mixed with West Slavs & a bit of German mixture in West Slavs.

I mean... I heard that 15% - 25% of Germans have Slavic orgins surnames.

So... I mean German mixture with Slavs is certainly a statistical number.

But... If Poland & Germany have gotten closer by DNA since the Iron age.

It seems to support that Poland & German were even further separate entities before the Iron age.

I suspect Germany was a Celtic Isolate at the time!? Perhaps!?

But.... It all seems to support that Slavs are the original inhabitants of Poland!!!

That even Polabians are the original inhabitants of Denmark!

Unknown said...

Another Question you should be asking yourself if Poland was a German territory then the Slavs moved in, then why is their so little Germanic Dna in Lithuania and vise versa? Surely they would have intermingled?
It looks like German Dna followed only during the well documented land grabs.

Davidski said...

Poles and Lithuanians appear to be native to the southern Baltic area. That's what ancient DNA suggests. For instance, see here...

So it's difficult to imagine that the ancient Wielbark and Przeworsk people who once lived in Poland were Germanic.

Germans by and large have origins in other parts of Europe, further to the south and west, and are more Middle Eastern genetically. But many Germans also do have Slavic admixture, and those from the former Eastern Prussia might have some Baltic Prussian ancestry.

By the way, no I don't think Huns and Avars were Slavs. They were most likely Turks, but they didn't leave much of a trace in Europe.

Unknown said...

You could be right,but realistically you need large numbers of men and women to service an army (or the locals would beat you back with sharp sticks) therefore you would think eventually they would leave a trace like the moors of Spain or the Saxon's in Britain.
In 1500 years time should we designate Bolsheviks and Mensheviks
as wild barbarian tribes invading Eastern Europe?

Davidski said...


Slavs weren't in perpetual conflict with Avars and Huns. Often they worked together because they were attacking and oppressing the same targets and not each other.

I think this pdf explains the situation nicely...,d.dGc&cad=rja

It seems there are many false cliches in the west about Eastern European prehistory and history, and it'll take years, if not decades, to put them to bed with ancient DNA and more advanced archeological work.

Unknown said...

Thanks for the link.
I was merely suggesting that because of the lack of Central Asian haplogroups (often suggested by historians as the homelands of avars and huns) and the similar movement's of the slav,ante, avar and Hun suggests that in fact they could have been one and the same.
Of course further research will dispel any ambiguities and help better understanding of eastern- europe which has be woefully neglected by scholars.

Can you give us any information when Y DNA will be tested from the ancient sites.

Lion heart said...

Davidski, I've seen u on many threads on the internet and u seem very knowledgeable about the genetics of West Slavs. Could u please maybe point out some very good studies about the genetic legacy of Slavs in Slovakia and even maybe genetic relationship between all west Slavs like poles and Slovaks? I seem to only find the same old genetics studies online and usually just about poles and Czechs with Slovakia always left out. As I'm 3/4 Slovak I would greatly appreciate it of u maybe knew some interesting articles or studies. Thank you

Davidski said...

Not sure about any useful papers on Slovaks, but they're very similar to both Czechs and Poles. They're basically like a cross between Hungarians and Poles.

But modern Hungarians are almost totally of West Slavic and German ancestry, so that doesn't mean Slovaks are part Ugric or anything.

Here is a PCA plot with Slovaks and Poles.

I've also run other tests with these Slovak samples, like here.

Lion heart said...

@Davidski, man thanks alot, the PCA plot is great. Would it be possible some how to fit my dad into that plot and maybe even me if I sent you our Raw data? I have it uploaded to Gedmatch. My dad is full slovak and ive traced his Family tree back to the early or mid 1800s and even to the late 1700s in the area right North of nitra. Maybe he could be a good sample of another Slovak for further studies of yours as his Background is basically all from one Region and seems very old there. Thanks again man