search this blog

Saturday, December 24, 2016

Do it right or don't do it at all

At least a couple of academic teams working on the population history of Europe need to read this PDF and take it to heart.

This article reviews scientific publications that have attempted to use genetic and genomic data in order to investigate European migrations between the fourth and ninth centuries. It considers early single-locus studies that used mtDNA and y-chromosome data. These studies were successful in formulating hypotheses concerning migration and heterogeneity, primarily between the Continent and the British Isles and Iceland, but could only examine a small portion of the entire genetic inheritance. The article continues with a presentation of more recent genome-wide studies. In particular, it evaluates the problems of using modern genomic data to understand past migratory processes, arguing that modern DNA is a problematic source for understanding population histories of the past fifteen hundred years and urges the sequencing and analysis of ancient DNA. It also presents some of the problems of re­search teams that did not include archaeologists and historians as integral participants in the planning, collection, and evaluation of data. It concludes with a brief outline of the authors’ current project that examines migration between Pannonia and Italy in the sixth century.


This work is necessarily interdisciplinary, something notably lacking in the majority of genetic studies. The Ralph and Coop study, while highly rigorous at the level of the population genetic analysis, included no historians or archaeologists, and the only historical literature cited, presumably to »identify« the Hunnic contribution to European population, was a general history of Europe, [25] a survey of Slavic history, [26] and two articles in the New Cambridge Medieval History. [27] The Busby et al. study also included no historians or archaeologists on its team, and the only historical literature cited was a Penguin History of the World, Peter Heather’s survey of the Early Middle Ages, and a survey of Muslims in Italy. [28] Unlike these studies, designed and executed exclusively by geneticists who then look through a few general historical handbooks to try to find stories that might explain their data, historians and archaeologists are integrated from the start in our project.

Patrick J. Geary and Krishna Veeramah, Mapping European Population Movement through Genomic Research [PDF download], Medieval Worlds, No. 4, 201, 65-78


Nirjhar007 said...

I think the paper is good, thanks . I will read it carefully. For the moment , I can say that Modern DNA research is very important still and will remain important . aDNA is still in its early stages and we will need more time .

It is of course necessary ,since Geneticists are working on historical and pre-historical aspects , a harmonious and continuous relationship, with Archaeologists and Historians is established, in most of the occasions .

And of course Merry 25th of December and sincere wishes to David and other folks here . The next time I say this , I know things will be changed a lot :) ...

Palacista said...

It would be important to add experts in historical linguistics to the list.

Matt said...


"These are large questions, and it will take many years and millions of euros to properly sequence skeletal remains from the over 1,200 graves in our sample. We must then analyze our data using statistical methodologies developed by population geneticists, and then confront the results with those developed from the analysis of stable isotopes, historical evidence, and cultural archaeology. Our project, although under way for four years, is still at the beginning.

However we hope, through close collaboration between disciplines and mutual respect for the contributions of historians, geneticists, and archaeologists, to begin to uncover the demographic history of the migration period and in so doing contribute to a map similar to that produced by Novembre et al., but of Europe’s population a millennium and a half ago. "

Good ideas - once you have a genetic summary map of Europe (and PCA might be the best way to do it), at one time period, you can compare to later time periods and so infer intensity of migrations. Very, very sampling intensive. Very expensive at the moment.

One thing we can do at the moment I think is, with the adna we have, for Europe is we can generate predictions of relatedness between populations based on inferred proportions of EHG, CHG, WHG, Anatolian, then we can compare that to actual relatedness on a Novembre PCA, or to Fst scores.

Having done similar things, using the West Eurasia and World PCA scores compared to European PCA, I tend to find some small signals there of excess relatedness between populations who share language groups and who are spatially related. It seems like the genetic maps of Europe adjusted for affinities on a world or west eurasia ancients scale are more geograpically and linguistically correlated and more structured in ways that mirror the y-dna as well, than they are without taking account of it.

Gioiello said...

Youns Azmi My RSRS & rCRS VALUES ! there is anyone here who can help me understand my Mt result?
Gioiello Tognoni Haplogroup H1.
Gioiello Tognoni It is an H1*, but it is waiting for its subclade, having some mutations: A3564G, T9110C. Also 573.1C. Very likely it is an old H1* of your country. Where do you come from?. Northern Africa perhaps?
Gioiello Tognoni Very likely you have these matches in Europe, thus I have no doubt that this hg expanded from Atlantic Europe to Northern Africa:
User ID HVR1 Mutations HVR2 Mutations
Q4WZN 519C 263G, 315.1C, 573.1C
7XTNS 519C 263G, 315.1C, 573.1C
8DVD6 519C 263G, 315.1C, 573.1C
KNXD5 519C 263G, 315.1C, 573.1C
KXYXK 519C 263G, 315.1C, 573.1C
Q787Q 188A, 519C 263G, 315.1C, 573.1C
R4RR8 136C, 519C 263G, 315.1C, 573.1C
u95k3 188A, 519C 263G, 315.1C, 573.1C
V4Y7P 189C, 519C 263G, 315.1C, 573.1C
Gioiello Tognoni Here you have a sample from Sardinia, where this hg may come from:
EU597532(Sardinia, Italy HGDP01077) Hartmann Haplogroup [H1] 06-APR-2008
T152C A263G 309.1C 315.1C A337G C522- A523- 573.1C A750G A1438G G3010A C3107N A4769G G5471A A7570G A8860G A15326G T16519C


J.S. said...

A Time Series of Prehistoric Mitochondrial DNA Reveals Western European Genetic Diversity Was Largely Established by the Bronze Age

A major unanswered question concerns the roles of continuity versus change in prehistoric Europe. For the first time, genetic samples of reasonable size taken at multiple time points are revealing piecemeal snapshots of European prehistory at different dates and places across the continent. Here, we pull these disparate datasets together to illustrate how human genetic variation has changed spatially and temporally in Europe from the Mesolithic through to the present day. Mitochondrial DNA (mtDNA) haplogroups were determined for 532 European individuals from four major eras: the Mesolithic, Neolithic, Chalco-lithic (late Neolithic/early Bronze Age transition) and Modern periods. The Mesolithic was characterized by low mtDNA diversity. These initial European settler haplogroups declined rapidly in the Neolithic, as farmers from the east introduced a new suite of mtDNA lineages into Western Europe. For the first time, we show that the Chalcolithic was also a time of substantial genetic change in Europe. However, rather than the arrival of new mtDNA lineages, this period was characterized by major fluctuations in the frequencies of existing haplogroups. Besides the expansion of haplogroup H, there were few major changes in mtDNA diversity from the Chalcolithic to modern times, thus suggesting that the basic profile of modern western European mtDNA diversity was largely established by the Bronze Age.
Discover the world's research

Samuel Andrews said...

^So everyone knows. That study didn't sample any new ancient mtDNA.

Matt said...

One comment I would make is that to around their comment that:

"Yet, nowhere in Novembre et. al.’s map can one find clear evidence of the migrations, the population exchanges, or the diffusions of more recent centuries, particularly those of the so-called »migration age« (fourth to ninth centuries of the Common Era) that we are accustomed to encountering in our historical texts as well as in our archaeological work. "

then I'm not sure this isn't partly due to the nature of PCA, which is to summarize genetic variation between all samples in orthogonal dimensions, largest to smallest. Some of the signals of genetic similiarity which relate European populations to one another may be orthogonal to the primary dimensions picked up by PCA in its low dimension (1-2).

On that note, a representation that does seem to pick up some signals of migration is Mathieson's Spacemix plot, here - (using Bradburd's Spacemix).

It's similar to the PCA plot which is next to it, but you can see in Europe, for Eastern European populations, there are wide elipses for each population connecting a very similarly positioned point at one terminus to a set of different end points. IRC this is how Mathieson has chosen to represent admixture edges. So in a sense looks to be representing a Slavic migration.

(OTOH, Novembre didn't find that when running Spacemix with just the Popres samples - That may be because with just Europeans alone, there was no need to add an extra admixture edge to represent the genetic covariance between samples).

That model by Mathieson included North Africans, who dominate the plot, so would be interesting to see how that would behave in a West Eurasia minus North Africans model.

Also, Merry Christmas ;) .

andrew said...

The case for interdisciplinary synthesis is a good one and often missing - the Chinese literature is particular weak in that regard.

Razib Khan said...

Is Nirjhar007 ever NOT the first commenter on posts on this blog?

Nirjhar007 said...

Razib, if you want me to follow your blog in order to raise its lousy followship, just say so...

Davidski said...

Nirjhar, since you don't follow my blog, how do you manage such a rapid response whenever I post something that piques your interest?

Nirjhar007 said...


MfA said...

I think you can follow a site anonymously on Blogger, under section "Manage Blogs I'm Following".

Davidski said...

Ahh, mystery solved.

Alberto said...

The paper is a good read, and the project is very interesting. But given the amount of time and money required for such extensive sampling I doubt it's doable in a larger scale, across space and time. But anyway I think we'll learn good things from it that will be applicable to other cases.

Isn't this team the one also looking at Globular Amphora Culture genomes? There was a strange title in that abstract about GAC that referred to Lombards, though I fail to see the relationship that GAC might have to this study.

Davidski said...

No, that was a mistake. Whoever made the abstracts PDF for that conference in Italy put a different title and author list with the abstract.

So the GAC paper is being done by a different team, from Italy and Poland, and should be out this week apparently, rumor has it.

Jim said...

"The article mentioned historians, archaeologists and linguistics. Those are not sciences, so why would any scientist want to associate those them. Think the caste system, those people are the Dalits."

That's a good list of the fallacies this caste system is built on.

1. Historians don't belong in with archeologists and linguists. It's a category error. Both archeology and linguistics are empirical sciences, along with zoology, botany, paleontology - I hope these people do believe in evolution - and geology. For that matter so is genetics; odd that this attitude would occur among geneticists.

2. If the objection to regarding them as sciences is that they are not based on experiment and falsifiable predictions, that is an error too. in the case of linguistics for instance there are numerous examples of linguists positing connections they cannot prove but that are later proven to be accurate and valid. The same is turning out to be true in paleontology. Another example in linguistics is Saussure's conjecture that vowel length in PIE was conditioned by a lost laryngeal. It was basically a guess. Not long after that Hittite texts were discovered and analyzed, and guess what they found - the lost laryngeal right where the sister proto-languages had long vowels.

3. The assertion that scientists should not work with non-scientists is sophomoric. Multiple lines of evidence lead to the most solid conclusions and sometimes that evidence comes from non-scientific disciplines, such as historiography.

But as for the existence of this caste system, you are completely right.

postneo said...

For gods sake !

We have heard "Saussure's conjecture" parroted so many times. Its over 100 years old. That too there's no common consensus among linguists on laryngeals.

Meanwhile other sciences march on with experiments, results and predictions at a blistering pace.

The truth is that linguistics unlike other sciences, has not evolved a basic non subjective formalism. Also It's data acquisition rate(field studies) is glacial and pathetic. So while intentions are good empiricism is not its strong suit.

Jim said...

"We have heard "Saussure's conjecture" parroted so many times. Its over 100 years old."

So is Newtonian physics, and Relativity too by now. Your point?

"That too there's no common consensus among linguists on laryngeals."

Just plain false, aside from fringe crackpots and Chomskyan theoreticians.

"The truth is that linguistics unlike other sciences, has not evolved a basic non subjective formalism."

The truth is that the formalism in matters of descent and genetic relations between languages that linguists developed in the 19th century was considered formal enough for paleontologists to adopt it. I suppose paleontology is not a science either and evolution is a satanic sham, right?

The "linguists" who lack a non-subjective formalism are the Chomskyans, hardly the whole field, and not considered real linguists outside the Anglosphere. You seem to be mistaken about what constitutes the field.

Descriptive linguistics is as empirical and formal as botany.

martin said...

Cela évitera d'affirmer que l'eushkara n'est pas une langue indo-européenne alors que les VRAIS spécialiste sont toujours dans l'indétermination.