search this blog

Wednesday, September 2, 2015

Finno-Ugric Poles in Kushniarevich et al. 2015

PLoS One has just put out a major paper on the genetic structure of Balto-Slavs, and unfortunately I have to say it's a major disappointment.

Kushniarevich A, Utevska O, Chuhryaeva M, Agdzhoyan A, Dibirova K, Uktveryte I, et al. (2015) Genetic Heritage of the Balto-Slavic Speaking Populations: A Synthesis of Autosomal, Mitochondrial and Y-Chromosomal Data. PLoS ONE 10(9):e0135820. doi:10.1371/journal.pone.0135820

The authors used a very small number of Polish samples for their genome-wide analysis, mostly from the Estonian Biocentre (see here and here). At least 10 of these Poles come from Estonia, and some even resemble northern Russians with their unusual ancestry proportions. Note the dichotomy in the levels of the lemon yellow "Siberian/Volga-region" component within the Polish set in the ADMIXTURE bar graph from the paper.

I raised this issue with Estonian Biocentre Research Director Mait Metspalu a while ago when these samples were first published, and this was his response.

What we have is self identity. As you can see from Supplementary table 1 these samples have been collected in Estonia and the donors are self reported Poles.

Well, I'm sure there are people in South America who identify as Spanish. But would anyone in their right mind use them to make inferences about the genetic structure and history of the people of Spain?

Fine-scale population genetic analyses like this should only be done with lots of samples from the right places. Self-reported Poles from Estonia, and perhaps also Russia, who clearly don't resemble Poles from Poland aren't good enough. Estonian Biocentre scientists do a lot of useful work, but in this particular instance they were rather sloppy in labeling these Estonian Poles as simply Polish.

Kushniarevich et al. were very sloppy in not stating where their Polish samples were really from (their map shows the middle of Poland) and not bothering to remove obvious outliers from their dataset.

See also...

Lipka Tatars vs Balto-Slavs

Recent admixture in West Eurasia (including Europe)


Dmytro said...

What exactly did they do with these samples in the YDNA columns? Did they ask the self-declared Poles to identify with one of the 6 regions into which they broke up the whole tested Polish group? I notice BTW that there are similar "outliers" in many other groupings (if one judges by the toothcomb projections...)

Davidski said...

The Polish Y-DNA samples were collected in several regions of Poland and belong to people who were native to those regions. But none of the groups were from significantly west of the Vistula, which means the Y-DNA data isn't representative of Poland.

I can't see any comparable genome-wide outliers in most of the other groups.

Dmytro said...

I don't understand. Aren't Wroclaw and Silesia (their Pol3)significantly West of the Vistula? Kashubia too, though not as much. They do indicate they took their YDNA samples from both West and East of the river (initial map). What I was curious about was which of these 6 Y DNA groups did they assign their "Ugro-Finn Poles" to. Or is that unclear?

As to the other "outliers" I was just going by the combprick extensions in other parts of the autosomal map. Many of them appear to be as sharp as the ones sticking out from the Polish group, but not so much for other Slavs. Which does appear inconsistent.

Davidski said...

Y-DNA data are indicated on the map by black dots and mtDNA data by white dots. Autosomal data points are marked (incorrectly for Poland) by triangles. The Estonian Polish samples didn't contribute any Y-DNA data to the analysis.

Most people in Wroclaw are descendants of Poles from former southeastern Poland, mostly from what is now Lviv in Ukraine. So this sample, at best, represents eastern Poland.

What this paper is missing are autosomal and Y-DNA data from places like Poznan, Bydgoszcz, Torun, Upper Silesia, and Krakow. In other words, regions located in pre-WWII Western Poland and other core areas of medieval Poland.