Thursday, January 26, 2017

Fine scale genetic affinities of Estonians and Finns (Haller et al. 2017)


Over at PLoS ONE at this LINK:

Abstract: Ancestry information at the individual level can be a valuable resource for personalized medicine, medical, demographical and history research, as well as for tracing back personal history. We report a new method for quantitatively determining personal genetic ancestry based on genome-wide data. Numerical ancestry component scores are assigned to individuals based on comparisons with reference populations. These comparisons are conducted with an existing analytical pipeline making use of genotype phasing, similarity matrix computation and our addition—multidimensional best fitting by MixFit. The method is demonstrated by studying Estonian and Finnish populations in geographical context. We show the main differences in the genetic composition of these otherwise close European populations and how they have influenced each other. The components of our analytical pipeline are freely available computer programs and scripts one of which was developed in house (available at: www.geenivaramu.ee/en/tools/mixfit).


Haller T, Leitsalu L, Fischer K, Nuotio M-L, Esko T, Boomsma DI, et al. (2017) MixFit: Methodology for Computing Ancestry-Related Genetic Scores at the Individual Level and Its Application to the Estonian and Finnish Population Studies. PLoS ONE 12(1): e0170325. doi:10.1371/journal.pone.0170325

32 comments:

  1. Dave, In general sense , how much Yamnaya type ancestry you think Estonians and Fins posses?.

    ReplyDelete
  2. About the same as their Indo-European speaking neighbors: 45-50%.

    But Finns usually don't show great fits with all of their steppe-related ancestry coming from steppe sources, which makes me think that they do have some extra EHG or SHG that other Northern Europeans, except a few like Saami and Northern Russians, basically lack.

    ReplyDelete
  3. Thanks. Previously Rob suggested an Iron age Uralicization of IEs, do you agree?.

    ReplyDelete
  4. Nirj

    I meant perhaps a few around the east Baltic, that is if aDNA confirms that steppe arrived before "Siberian". Too many holes still to confirm

    ReplyDelete
  5. It seems like the Siberian/East Eurasian component that we associate with almost all Uralics, except Hungarians, arrived in and around what is now Estonia during the Late Bronze Age or early Iron Age, and that's probably also when Uralic languages spread to the East Baltic.

    But considering the Corded Ware-like genetic structure of Estonians, it doesn't seem like there was anything like a total population replacement in the region since the Corded Ware period.

    ReplyDelete
  6. Thank you guys, yes I agree . More continuity through genes but languages did change .

    ReplyDelete
  7. Valter Lang from University of Tartu has a recent theory about the linguistic situation:

    "A common horizon for both Pre-Proto-Saami and Pre-Proto-Finnic (if it indeed existed) can only be found in the Volga–Oka region in the Bronze Age;

    From there two branches of cultural influences spread westwards, one through the North-Western and another through the South-Western Passage of Contacts. The former can be connected with Pre-Proto-Saami and the latter with Pre-Proto-Finnic.

    Both movements took place in several waves lasting over many centuries. The first Pre-Proto-Saami movements perhaps started already within the Textile Ware networks, but they certainly continued in the later Bronze Age / Early Iron Age (Anan’ino influences). The Pre-Proto-Finnic speakers started to shift westwards at the end of the second millennium BC.

    ...

    As a result of these processes and language contacts with Proto-Baltic, Proto-Germanic, and some Palaeo-European, Proto-Finnic emerged and it also achieved the dominant position at least in what are today coastal Estonia, SW Finland, and the Daugava valley in Latvia. However, as more intensive and developing processes concentrated next on the coastal areas further north, it is easy to imagine the mechanisms of the separation of one portion of Proto-Finnic-population – the one that was later called South Estonian. "

    https://s12.postimg.io/m49gshl71/image.png

    http://oi63.tinypic.com/15g80nq.jpg

    "Archaic" South Estonian Lang mentions is the same as Voro-Seto here, it's spoken in Southeastern Estonia.
    http://oi63.tinypic.com/14txao0.jpg

    North Finnic (Finnish, Karelian, and Vepsian) is a young split, and expand into Saami area, which would explain their eastern shift towards modern Saami in particular relative to other Finnic speakers.
    http://evolbio.ut.ee/CGgenomes_analyses/PCA/PC1_PC6_for_DiversitySetM5LDP10005002MEDIANS_wh_8_x_8_inches_cex=0.5.pdf

    The Paleo-Lakelandic/Lapplandic people are possibly related to the Bolshoy Oleni Ostrov people, but that can't be resolved without ancient DNA.

    ReplyDelete
  8. @ Shaikorth

    I thought "An essay on Saami ethnolinguistic prehistory" by Aikiko was pretty good.

    ReplyDelete
  9. @Rob

    Yep, read it. What Aikio lays out there is also compatible with Lang's scheme, he locates the original separation of pre-proto-Saami in Volga-Oka region and so on. The issue with population replacement in Eastern Baltic could be a tricky one. The proposed southern passage is well within the horizon of Corded Ware and derivatives so without aDNA it's hard to say how different the locals and any LBA arrivals were.

    ReplyDelete
  10. Shaikorth
    Do you have the full title of the Lang ?

    ReplyDelete
  11. @Rob


    Formation of Proto-Finnic – an archaeological scenario from the
    Bronze Age / Early Iron Age

    One issue (though not too relevant to the main scheme) I found with the paper is that Lang mentions a "proto-Germanic" presence very early in the Baltic while the loanword evidence points to an earlier or parallel stage of development not showing Grimm's Law. More on that in:
    The Prehistoric Germanic Loanword Strata in Finnic

    ReplyDelete
  12. Does that mean the Siberian/East Eurasian DNA and even possibly Uralic does not correlate to N1c? N seems to have been in Europe longer than that given the current yfull dates and modern distribution. We know N was present 2500BC in west Russia which would fit well with the sample belonging to L1026.

    ReplyDelete
  13. Good work. The correlation between those results and regional ydna in Finland would be interesting.

    One obvious drawback. They suggested that Latvian components 9n Estonia implement higher diversity. It is not true in case two small country share same ancestry. It is quite a same idea as suggesting genetic diversity between Tallinn and Tarti in Estonia. It would be true if for instance Poland and Germany share common ancestry.

    ReplyDelete
  14. @JohnHutchins12

    Difficult to say without more ancient DNA from the forest zone. What were the Bronze Age populations of Volga-Kama and Oka like, was the pre-Uralic pre-Sami population of Karelia or Lappland Siberian-influenced like Bronze Age mtDNA from Kola suggests?

    In the Eastern Baltic I'd expect L1026 to correlate with Uralic, because the Y-full MRCA's there are Late Bronze Age, corresponding to Lang's proposal.

    ReplyDelete
    Replies
    1. I agree, more forest zone DNA is needed. I was hopeful the Baltic DNA paper would shed some light but it seems more focused on corded ware's influences on the region.

      Another thing about L1026 is that it's 4800ybp tmrca correlates well with the main branches of R1a and R1b's expansion, specifically corded ware. This would mean Uralic and Indo European expanded at the same time into North Eastern Europe. I am doubtful of this.

      Delete
  15. They are disappointingly opaque about how their MIXFIT is actually done.

    ReplyDelete
  16. @JohnHutchins12

    L1026 is a Corded Ware contemporary, but the branches found around the Baltic Sea presently are subclades of VL29 and Z1925, their MRCA's are 3600 ybp and 3000 ybp. This is more like a LBA arrival. In the Volga region there's older L1026 and even Mesolithic splits like Y9022, but those didn't reach the Baltic.

    ReplyDelete
  17. Yes but VL29 and Z1925 are almost exclusivly found around the Baltic. Early branches of Z1925 have also been found in Sweden. In my opinion these two clades didn't spread to the Baltic but were born there from a L1026 people who migrated to the Baltic earlier from West Russia.

    ReplyDelete
  18. So, basically:

    (1) Finns has a much larger "indigenous Finnish" component than the "indigenous Estonian" component in Estonians,

    (2) Finns receive a significant Danish contribution,

    (3) Estonians receive a significant Slavic contribution, and

    (4) both receive trace amounts of other European ancestry sources.

    ReplyDelete
  19. @JohnHutchins12

    VL29 and Z1925 are both quite a ways down from L1026 though. The former is 8 SNPs down and estimated at 3600 (3000-4300) years old according to Y-Full, while the latter is 19 SNPs down and 3000 (2500-3600) years old. So I agree with Shaikorth, the expansion of L1026 does not have to reach the Baltic immediately but can get there quite a bit later.

    Hadn't looked at Y-Full N tree for a while - it seems that VL29 and the Chukchi-Mongol B197/Y16323 branch are now united by the Y6058 SNP.

    The sister branch of VL29 shows up in a Komi sample. Under Z1936 we already knew that L1024 and its cousins could be found in Ob-Ugrics, Tatars, and Bashkirs. Z1934(xZ1925) seems to be scattered across European Russia and into Finland also.

    Even VL29 can be found at 13% in Mari and 5% in Samoyeds, though perhaps it's of recent Russian origin.

    ReplyDelete
  20. @capra internetensis

    Look at VL29 and Z1925 on Yfull. All of the early branches are from Finland, Sweden and one from America. This suggests that VL29 and Z1925 both originated somewhere near these two regions, I would suggest either SW Finland or Estonia. Vl29 is not 3600 years old that is its TMRCA, VL29 is actually 4200 years old, If we are to believe that VL29 was born in the Baltic and L1026 was not, these L1026 people migrated there between 4800ybp and 4200ybp. That puts its migration in the same time frame as Corded ware. Number of SNPs does not affect the dates if they are already given.

    Z1936 dose not affect VL29 at all, it simply had the same origin with L1026 in west Russia before splitting and going separate paths.

    Also, I believe the VL29 in Mari is N-Y10931 which downstream.

    ReplyDelete
  21. @JohnHutchins12

    The subclades of VL29 being Baltic implies that the MRCA of VL29 was in the Baltic, and the MRCA dates to the TMRCA, not the 'formed' date. The 'formed' date is the TMRCA of CTS10760.

    Thanks for the Mari info.

    ReplyDelete
  22. Yes, but the time between the formed date and TMRCA are usually when populations are not migrating meaning they occur in relatively the same area. This is why you see rapidly migrating successful lineages have the same formed and TMRCA date and dormant lineages having a formed date and TMRCA differing by thousands of years.

    ReplyDelete
  23. @JohnHutchins12

    If a lineage is migrating into new territory because the population is expanding rapidly, then there'll be lots of levels with TMRCAs close together. If they are migrating to seek a new opportunity there may be no growth until they get there. If it is a small and scattered population the lineage could coalescence anywhere with few surviving branches. If they are fleeing from a disaster they could be bottlenecked. I don't see any evidence that there's a general rule here.

    ReplyDelete
  24. OT, but of some interest. A new paper comparing endogenous DNA from tooth cementum and petrous bones. They tried with cremated samples too, and apparently there's not much hope to ever get DNA out of them:

    "We were not able to identify any authentic ancient human DNA in the cremated petrous bones from the Danish Iron Age skeletons. Clearly, the heat during cremation has fragmented the DNA beyond analytical recovery. Admittedly, we only have seven cremated bones in this analyses which are too few to definitively reject this possibility. Furthermore, cremation practices (including the degree of thermal damage) were different between cultures and time periods. However, unless contradicting results emerge in the future, our findings indicate that it is pointless to sample petrous bones from cremated skeletons for the purpose of genomic analyses. In turn, such samples have been shown to be useful in yielding strontium isotope ratios."

    http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0170940

    ReplyDelete
  25. Yep, that paper has new samples from the Central Asian steppes; probably Andronovo from northern Kazakhstan.

    Data Availability: All relevant data underlying this study and its findings have been uploaded to the European Nucleotide Archive with accession numbers PRJEB18722/ERP020675.

    I e-mailed Morten Allentoft, and the genotypes aren't available, since it wasn't that kind of paper, but let's see what happens when these reads go online.

    By the way, also...

    http://www.ebi.ac.uk/ena/data/view/SAMN04633627

    New EBA sample from Poland. As far as I know, he belongs to R1a.

    ReplyDelete
  26. Seeing the age of the dude ~2000 BC, should be one of Z-280s .

    ReplyDelete
  27. I've read Udmurts have the most Yamnaya ancestry. They speak a Uralic language( so I assume the IE folks were gone and replaced in some areas with Uralic folks?) DOES ANYONE KNOW IF, the mtDNA H6 sample from the ANDRONOVO CULTURE was the full sequence? thank you, jv

    ReplyDelete
  28. @capra internetensis

    You are correct that it is not a general rule but a trend in population genetics.

    Just look at the proposed dates of the Indo European Migrations and the R1a tree on YFull. M417 is formed 6500BC living in relatively the same area in the steppes until its TMRCA of 3500BC which is when the large scale IE migration occurred.

    L1026 may not have been in the Baltic as early as I have suggested but I would definitely not rule it out until we have ancient DNA.

    ReplyDelete
  29. @JohnHutchins12

    I surely wouldn't rule it out either, hopefully we will get that Baltic aDNA paper soon and find out.

    ReplyDelete
  30. 61.5% N in Finland
    42% Lithuania
    38% Latvia
    34% Estonia
    23% Russia
    10% Belarus
    Rest of Europe tiddly squeaks or less.

    N clearly ebbed and flowed across the Eurasian snowline well into the post LGM melt (the line is very high).

    I really like Andrews summary of the paper. I would add that it looks like southern Finns travelled to Estonia but not the other way around in any numbers. An SOUTH EASTERN movement. Was this a retreat from the snow or evidence of the main direction of flow?

    In this region it is the story of N that is most interesting. Was this a migration from/to asia, a trade/raid route, a Reindeer run or just another Y haplogroup boom/bust? Just another example of Y haplogroup instability in populations. The story of N still feels mostly untold. Which was disappointing.

    ReplyDelete

Read the rules before posting.

Comments by people with the nick "Unknown" are no longer allowed.

See also...


New rules for comments

Banned commentators list