Negligible genetic flow in Slavic expansion to the Balcans

A new genetic study comes to confirm what most of us already knew: that Southern Slavs don't show any significant signature of immigration from the core Slavic area North and NE of the Carpathian Mountains that can be attributed to the so-called Slavic migrations of the Dark Age.

Alena Kushniarevich et al., Genetic Heritage of the Balto-Slavic Speaking Populations: A Synthesis of Autosomal, Mitochondrial and Y-Chromosomal Data. PLoS ONE 2015. Open access → LINK [doi:10.1371/journal.pone.0135820]

Abstract

The Slavic branch of the Balto-Slavic sub-family of Indo-European languages underwent rapid divergence as a result of the spatial expansion of its speakers from Central-East Europe, in early medieval times. This expansion–mainly to East Europe and the northern Balkans–resulted in the incorporation of genetic components from numerous autochthonous populations into the Slavic gene pools. Here, we characterize genetic variation in all extant ethnic groups speaking Balto-Slavic languages by analyzing mitochondrial DNA (n = 6,876), Y-chromosomes (n = 6,079) and genome-wide SNP profiles (n = 296), within the context of other European populations. We also reassess the phylogeny of Slavic languages within the Balto-Slavic branch of Indo-European. We find that genetic distances among Balto-Slavic populations, based on autosomal and Y-chromosomal loci, show a high correlation (0.9) both with each other and with geography, but a slightly lower correlation (0.7) with mitochondrial DNA and linguistic affiliation. The data suggest that genetic diversity of the present-day Slavs was predominantly shaped in situ, and we detect two different substrata: ‘central-east European’ for West and East Slavs, and ‘south-east European’ for South Slavs. A pattern of distribution of segments identical by descent between groups of East-West and South Slavs suggests shared ancestry or a modest gene flow between those two groups, which might derive from the historic spread of Slavic people.

This is most evident in the identity-by-descent (IBD) analysis:

Fig 4. Distribution of the average number of IBD segments between groups of East-West Slavs (a), South Slavs (b), and their respective geographic neighbors.

The x-axis indicates ten classes of IBD segment length (in cM); the y-axis indicates the average number of shared IBD segments per pair of individuals within each length class.

For non-acquainted: shorter segments (left) indicates older relatedness, now very fragmented by repeated chromosome recombination, while longer segments (right) indicate more recent one, which had less time to be chopped into pieces.



The authors explain:

The presence of two distinct genetic substrata in the genomes of East-West and South Slavs would imply cultural assimilation of indigenous populations by bearers of Slavic languages as a major mechanism of the spread of Slavic languages to the Balkan Peninsula. Yet, it is worthwhile to add here evidence from the analysis of IBD segments: the majority of Slavs from Central-East Europe (West and East) share as many IBD segments with the South Slavs in the Balkan Peninsula as they share with non-Slavic populations residing nowadays between Slavs (Fig 4A and 4B; Table G in S1 File). This even mode of IBD sharing might suggest shared ancestry/gene flow across the wide area and physical boundaries such as the Carpathian Mountains, including the present-day Finno-Ugric-speaking Hungarians, Romance-speaking Romanians and Turkic-speaking Gagauz. A slight peak at 2–3 cM in the distribution of shared IBD segments between East-West and South Slavs (Fig 4A and 4B) might hint at shared “Slavonic-time” ancestry, but this question requires further investigation.

Another graph of interest is surely the Principal Component Analyses of the three types of genetic markers:

Fig 2. Genetic structure of the Balto-Slavic populations within a European context according to the three genetic systems.

a) PC1vsPC3 plot based on autosomal SNPs (PC1 = 0.53; PC3 = 0.26); b) MDS based on NRY data (stress = 0.13); c) MDS based on mtDNA data (stress = 0.20). We focus on PC1vsPC3 because PC2 (S1 Fig) whilst differentiating the Volga region populations from the rest of Europeans had a low efficiency in detecting differences among the Balto-Slavic populations–the primary focus of this work.

In the mtDNA graph (c) it is hard to discern any pattern, as the various studied populations seems to form rings of eccentricity around the Balcans, probably because no Western Europeans are present in this particular PCA. 

However in the autosomal (a) and Y-DNA (b) figures more defined patterns do emerge. Quite apparently in all three graphs, South Slavs appear as strictly Balcanic. 

More interesting is probably the relative position of Russian and Baltic speakers: the first showing very notable diversity almost representative of the whole East European region and again indicative of assimilation rather than replacement being the main drive in Russian ethnic expansion, at least in the North. 

Balto-Slavic peoples appear intermediate between Russians and Finns (and overlapping Estonians) in the Y-DNA graph and somewhat extreme in the autosomal graph, something that comes as no surprise, as they seem the best preserved vessel of Eastern Paleoeuropeans. Curiously a few Sorbian individuals also tend to that same extreme, what may well be a reason to increase interest on the study of this forgotten and neglected Slavic minority of Eastern Germany. Their Y-DNA is, also intriguingly, most similar to that of Swedes, rather than to their geographic neighbors or ethno-linguistic relatives.

Other Western Slavs, form two clear distinct sub-clusters: with Czechs being notably more Western than Poles and Slovaks, who tend to cluster with mainline Russians and Ukrainians instead. One can of course think that this Polish-Slovak-Ukranian-Russian cluster could be the demic or genetic core of the Slavic cluster. However I can't but wonder how much of that clustering, as well as the differences shown by Czechs and Sorbians should be attributed to older periods like those of Corded Ware Culture, Eastern Bell Beaker, etc.

Mitochondrial DNA of some Slavic peoples

With emphasis on haplogroup H5 and H6.

Marta Mielnik-Sikorska et al., The History of Slavs Inferred from Complete Mitochondrial Genome Sequences. PLoS ONE 2013. Open access → LINK [doi:10.1371/journal.pone.0054360]

Most interesting is maybe table 1 (right), which lists two Polish populations (Kashubia, at the Baltic coast, and Podhale: the Carpathian piedmont), Ukrainians and Czechs.

We can see here that the most common lineages among these Slavs are not different from other European populations, namely H*, H1, U5a, U5b and also the, arguably Neolithic, lineages J1 and T2. I find relevant in this sense that there is a significant amount of T(xT1,T2) among Kashubian Polish especially.

Another point of interest is the minor presence of North and Central Asian lineages A, C, D and G, for which the authors present an elaborate rationale:

... we were able to pinpoint some lineages which could possibly reflect the relatively recent contacts of Slavs with nomadic Altaic peoples (C4a1a, G2a, D5a2a1a1).

They also suggest that the L2a1l2a, found among the Polish, is of Ashkenazi Jewish origin. L1b1a8 found in Polish and Russians belongs to the wider L1b1a, recently argued to be European-specific.

Another point of notice may be the rare HV0(xV) found at significant frequencies among Ukranians (4.5%).

But the authors make a particular effort to discern within haplogroups H5 and H6, which they find of particular interest. H5 might be (with doubts) of Italian origin and they consider its coalescence age (on the dubious molecular clock estimate methods) as clearly pre-Neolithic.

Based on these speculative methods they argue that several Slavic-specific clades within H5 may be contemporary in origin with U4a2, common in Central and Eastern Europe. They consider both to be roughly from Early Neolithic times.

Figure 1. Complete mtDNA phylogenetic tree of haplogroup H5.

Legend:

→ Green: Polish, Czech, Slovaks, Ukrainians and Russians
→ Red: German, Dutch and Austrians

→ Yellow: Italians and Spaniards

→ Blue: Irish, British, Danes and Finns

→ Magenta: Tunisia

→ Black: Levant

→ Grey: USA

→ White: unknown geography

If they are correct in their interpretation of the tempo of H5, the hypothesis of H sublineages migrating Northwards, from Southern to Central Europe, within Neolithic would seem to gain some support.

However H5 is less common in the Czech Republic and Austria than in Poland or Ukraine and the Neolithic colonization of Poland should have gone via the Czech Republic and, previously, Austria. Of course we cannot reject upfront a founder effect specific to Poland but what about Ukraine, which was almost totally oblivious to the Balcano-Danubian Neolithic phenomenon?

The other focus is H6, which is found almost only among Ukranian and Czechs of the four target populations. Generally speaking H6a and the most rare H6c are European, while H6b is Central and West Asian. In spite of its extreme rarity, the authors detected H6c in three individuals (one Czech, one Pole and one Slovak), all non-Slavic H6c are from Central-West or NW Europe (or from unknown locations). This seems to define H6c as a rare Northern-Central European haplogroup (excluding Eastern Europe apparently).

The remaining H6 samples sequenced in our study belong to different H6a subclusters being identified as singletons (H6a1a*) or as members of subclusters H6a1a4, H6a1a9 and H6a1b3. Subcluster H6a1a9 is novel, comprising of two haplotypes found in Russians and Ukrainians. Subcluster H6a1b3 is also interesting because it contains, except for European individuals of unknown origin, a founder haplotype of Czech origin and two Polish haplotypes.

Figure 2. Complete mtDNA phylogenetic tree of haplogroup H6 (legend as above).