Pankratov (2020): Differences in fine level local population history in Estonia

Open access paper Differences in local population history at the finest level: the case of the Estonian population, by Pankratov et al. EJHG (2020).

Interesting excerpts (emphasis mine):

The PCA shows the presence of a genetic gradient within Estonia with the main differentiation observed between South-East and North-East of the country in agreement with previous studies [2, 11, 12]. This differentiation reflects a broader-scale South-North gradient in Eastern Europe (Fig. 1b) with Estonians from the North-East being closer to Finns while South-East Estonians projected closer to Latvians and Lithuanians.

a Principle component analysis of the Estonian dataset. The first two PCs are shown. Individual dots are coloured according to the donor’s place of birth. Estonian counties were divided into four groups (SE South-East; SW South-West; NW North-West; NE North-East) as shown in the map. This map was created in R (https://www.R-project.org/) [16] using an shp object of the administrative and settlement units provided by the Estonian Land Board, 2018.11.01 (https://geoportaal.maaamet.ee/eng/Spatial-Data/Administrative-and-Settlement-Division-p312.html). See “Methods” for more details. The individuals with no information available regarding their place of birth are shown in grey. b Projecting Estonian samples onto PC space defined by European samples (“Methods”, Supplementary text section 1). Red crosses correspond to medians of European populations while empty circles represent individual samples. Populations are labelled as follows: Ita Italians; Spa Spaniards; Fre French; Ger Germans; Hun Hungarians; Eng British; Swe Swedes; Ukr Ukrainians; Bel Belarusians; RuCS Russians from Central and Southern Russia; Pol Poles; Lit Lithuanians; Lat Latvians; Mor Mordvins; RuN Russians from Northern Russia, Estonian samples are shown in colour reflecting their position along PC1 in (a). In both panels percentages in the axis labels show the proportion of the total variance explained by the corresponding PC.

In order to understand how gene flow barriers and/or differences in local population density shaped the IBD-sharing pattern in the R50+ dataset, we inferred migration surfaces using MAPS [16]. We used two windows of IBD segments length (in centimorgans (cM)), 2–6 cM and more than 6 cM, which under a simplistic model of infinite population size have mean segment ages of 50 and 12.5 generations, respectively [16]. The results for the two length bins generally agree with each other, suggesting higher levels of gene flow in the North along with a barrier separating South-East Estonia (Supplementary text 2.4). A second barrier, separating the islands, especially Hiiumaa, from the mainland is also evident. This observation suggests that the population ancestral to modern South-East Estonians was partially isolated from the rest of the country at least since 50 generations ago. Interestingly, this genetic differentiation is consistent with linguistic data suggesting that the deepest split within the Finnic languages separates Southern Estonian from the other branches of the phylum that includes Northern Estonian [17].

Modelled relative ancestral proportions of «Balts» (Latvians and Lithuanians), «Slavs» (Belarusians, Poles, Russians, Ukrainians), Finns, and Swedes attributed by applying non-negative least-squares approach (NNLS) to CHROMOPAINTER/fineSTRUCTURE (CP/FS) results are shown. See Supplementary text section 3.1 for details. The colour of each parish reflects mean values of samples coming from this parish. Parishes with no samples in the R50+ dataset are filled with grey. Names in rectangles show directions to neighbouring countries. These maps were created in R (https://www.R-project.org/) [38] using an shp object of the administrative and settlement units provided by the Estonian Land Board, 2018.11.01 (https://geoportaal.maaamet.ee/eng/Spatial-Data/Administrative-and-Settlement-Division-p312.html). See “Methods” for more details.