the russian journal of genetic genealogy european prehistory in mirror of genetics a contemporary...

8/14/2019 The Russian Journal of Genetic Genealogy European Prehistory in Mirror of Genetics a Contemporary View

1/12

RJGGThe Russian Journal of Genetic Genealogy: Vol 1,1, 2009

5

ISSN: 1920-2997 http://rjgg.org All rights reserved

European prehistory in mirrorof genetics: a contemporaryview*

Oleg Balanovsky

Received: June 8 2009; accepted: June 8 2009; published June 30 2009Research Centre for Medical Genetics, Russian Academy of Medical Sciences,

Moscow, RussiaThis study has been originally published in: Balanovsky O.P. (2009) Human

genetics and Neolithic dispersal' The East European Plain on the eve ofagriculture edited by Pavel M. Dolukhanov, Graeme R. Sarson and Anvar M.Shukurov. BAR -S1964. ISBN 978 1 4073 0447 2. P:235-46

1

Introduction

Numerous branches of knowledge are cur-rently affected by a certain eurocentrism, this be-ing especially the case of the studies focused onhuman population genetics. These studies were(and remain) developed predominantly in Euro-pean laboratories and are less popular on othercontinents. Population geneticists need humanpopulations as subject matter of their research,

and in most cases we study those which are inproximity of our homes. That is why Europe isthe continent which is by far better studied ge-netically, and the European genetic landscapeis much deeper understood and discussed thanany other part of the world. For the same reasonthe controversies between different schools ofthought regarding various aspects of the Euro-pean gene pool became much more apparent

The two concepts

The variation of "classical" genetic markers(which became referred to like that when theDNA came to the fore to replace them) was bestsummarized in the book by Luigi Luca Cavalli-Sforza and his colleagues published in 1994.Unsurprisingly, the largest chapter of this bookis the "European" one, which describes how theEuropean genetic landscape had been formedduring the Neolithic expansion from the NearEast. It was one of the most minutely-elaborat-

ed concepts in population genetics at that time;nonetheless it was almost entirely rejected inthe subsequent decade.

The European genetic landscape, as re-stored based on the analysis of classical mark-ers, shows three principle features: 1) a generalhomogeneity (the Europeans are geneticallyvery similar to each other, compared to popula-tions of other continents); 2) the presence ofonly a few outliers (isolated peripheral popula-tions such as Icelanders, Saami, or Sardinians);their peculiarities are the secondary, havingarose after these populations were demographi-cally split off and underwent the genetic driftfrom the main European corpus; 3) Clear geo-graphic patterns of gradual genetic changes.

To identify these geographic patterns Cav-alli-Sforza and his colleagues (Menozzi et al.,1978, Cavalli-Sforza et al., 1994) and indepen-dently Russian geneticists (Rychkov & Balanovs-kaya, 1992) developed the method of "syntheticmaps". These maps are created by a complex

mathematical algorithm but in a simpler waythey consist of displaying the geographic distri-bution of an "ideal" genetic marker, which cor-relates with geographical patterns of the major-ity of real markers presenting the data (Menozziet al., 1978; Rychkov & Balanovskaya, 1992;Balanovskaya & Nurbaev, 1997a). This synthet-ic map visually demonstrated gradual changeswith a remarkable geographical pattern: fromAnatolia via the Balkans over the rest of Europei.e. from the Southeast to the Northwest (Fig 1).This picture was interpreted as a result of the

gradual spread of farming (and farmers) acrossEurope which was known since Gordon Childe(1928) to follow the same trajectory.

This concept was additionally substanti-

1

*


2/12


6


ated in two ways. First, the "isogenes" (linesconnecting the same gene pools on the genet-ic map) have shown a remarkable agreementwith isochrones (lines showing the early arrivalof agriculture based on archaeological and ra-diometric evidence). Second, the concept andthe mathematical model of the so-called demicdiffusion was developed (Ammerman & Cavalli-Sforza, 1984). It implies a slow (generation bygeneration) migration of farmers which assimi-

lated indigenous populations, and thereby grad-ually dissolved the initial "farming" gene pool.As a consequence, the geographic trajectory ofmigration becomes a geographic line of gradualgenetic changes: from a "mainly farming" genepool in Anatolia to a "mainly indigenous" one inthe Europes north-west and north-east (as themost distant from Anatolia).

This elegant, reasonable and sufficientlysubstantiated concept of the origin and com-position of the European gene pool dominatedpopulation genetics in the 1980-1990s. Gener-

ally, four elements of this concept could be po-tentially criticised:i) the data-set (classical markers);ii) the methodology (synthetic maps);iii) the logical foundation (attributing south-

east-northwest pattern to Neolithisation) oriv) controversial results obtained with a use

of independent data, methods and logics. Crit-ics used all four elements but with a varaiblesuccess.

The popular idea that classical markers are"worse" than new DNA markers has never beenpositively proven and should be considered rath-er as a scientific fashion. Yet some critics tendto reject the classical markers arguing that theyare affected by natural selection, and therefore

their variation would be the result of both his-torical and biological factors. However, manyDNA markers can be equally affected by biologi-cal factors and therefore geographic distribution

of a genetic marker reflects the history which tosome degree was blurred by biological selectionaffecting this marker. And, second, the biologi-cal factors differently affect various markers andtherefore disappear when averaging, in the casewhen numerous markers are considered (Ya-mazaki & Maryama, 1973; Lewontin & Krakauer,1975; Balanovskaya & Nurbaev, 1997b).

The method of synthetic maps was at-tacked by Robert Sokal, who used an alternativemethod (autocorrelation analysis) for reveal-ing geographical patterns in the genetic data

(Sokal, Oden, 1978). Using computer simula-tion he demonstrated that synthetic maps com-piled from interpolated maps produce gradualpattern even from randomly permutated data,hence the obtained patterns are artificial (Sokalet al., 1999). However, our recent simulations(Balanovsky et al., 2008) failed to recogniseany difference between synthetic maps frominterpolated surfaces (criticized by Sokal andcolleagues), on the one hand, and from non-interpolated raw data (considered as control bySokal and colleagues), on the other. It is equallyremarkable that Sokal and colleagues did not

express doubt that the gradual genetic patternfrom Anatolia is the main feature of the Europe-an gene pool, yet they did question the methodsapplied for to identify this pattern.

Curiously enough, the applied logic conclu-sion (attributing the observed genetic patternto the Neolithic expansion as the both followedthe same trajectory) had never been criticizedto the best of our knowledge, though populationgeneticists were well aware that a correlationnever proves the cause-effect relationship. Theinterpretation in terms of the Neolithic expan-

sion seemed so obvious, transparent, and natu-ral, that this logical mistake became only appar-ent when controversial results started emergingfrom independent data.

The evidence, demonstrating the Palaeoli-thic time for the origin of the European genepool was based on mitochondrial DNA (mtDNA).The principal difference between mtDNA andY chromosomal markers on the one hand, andthe classical and autosomal DNA markers onthe other, resides in the presence or absence ofrecombination. Autosomal markers recombineand therefore each marker is inherited indepen-dently from all other markers. MtDNA and themain portion of the Y chromosome do not re-combine. That is why every occurring mutation

Figure 1. Synthetic map, summarizing genetic variation in

Europe (from Cavalli-Sforza et al., 1994).


3/12


7


gets transmitted from generation to generationalongside other mutations, which did occur ear-lier. In other words, the mutation (a mistake inthe genetic text) became forever part of this textand is transmitted together with all the mistakesthat had appeared in this text earlier. Whencomparing different texts (so-called haplotypes)it is possible to trace mutations back in time andreconstruct a "genealogy" of these texts. i.e. to

draw their "family tree". This tree is commonlyrooted in the most recent common ancestor (a"mitochondrial Eve") and each branch of thetree differs by its particular set of mutations.Next, each twig of a certain branch carries allmutations characteristic to this branch, togeth-er with a set of additional "twig-specific" muta-tions. These branches and twigs are called hap-logroups (subhaplogroups) each of them unitesa group of closely related haplotypes (which canbe compared with a leaves of this tree). Assum-ing an average rate of mutations one can calcu-

late the age of each haplogroup by multiplyingthe number of accumulated mutations by themutation rate.

This methodology, applied to the Europeanmitochondrial pool (Richards et al., 1996), dem-onstrated that most branches (haplogroups)found in Europe were much older that the Neo-lithic and most of them fell into the age range ofthe Upper Palaeolithic. Based on this evidenceit was concluded, that European gene pool wasformed by the initial peopling of the continentby anatomically modern humans (AMH) duringthe Upper Palaeolithic, and that it is still pres-ent in the most of present-day Europeans. Asfor the Neolithic expansion, it had, therefore,a limited impact on the European gene pool.

Hence, this new concept imposed the "culturaldiffusion" model of Neolithisation in contrast to"demic diffusion" model advanced by Ammer-man and Cavalli-Sforza (1984).

The following decade witnessed a heateddebate between two camps of geneticists, name-ly the "cultural diffusionists" and the "demists".Despite the ongoing debate, the methodologicallimits and benefits of both models are apparent.

The source database for demic diffusion modelwas much richer (hundreds of markers studiedin dozens of populations) while Richards and col-leagues were restricted to one marker (mtDNA)studied in a limited set of populations. Arguably,as Barbujani and colleagues (1998) pointed out,the Palaeolithic origins of haplogroups found inEuropeans do not necessarily imply that thesehaplogroups were present in Europe since thePalaeolithic. As haplogroups age is calculatedbased on its diversity, they could have accumu-lated diversity in other parts of the world ar-

riving into Europe being already diverse. Thisproblem of the pre-existing diversity met ele-gant solution in the following paper by Richardsand colleagues (2000), which became the mostrecognized study of European genetics. In thatpaper the founder mtDNA lineages were identi-fied which were deemed as the starting pointsfor the entire European diversity accumulatedin situ. Although different criteria for "founding"resulted in slightly different time assessments,all calculations demonstrated the Upper Pa-laeolithic age for most European clusters of lin-eages (haplogroups), while haplogroups whoseappearance in Europe can be attributed to theNeolithic period make up only a quarter of thetotal European gene pool (Fig. 2).

0 10,000 20,000 30,000 40,000 50,000

U: 5.7%

HV: 5.4%

I: 1.7%

U4: 3.0%

H: 37.7%

H-16304: 3.9%

T: 2.2%

K: 4.6%

T2: 2.9%

J: 6.1%

T1: 2.2%

Years before present

Neolithic

Mesolithic

LUP NUP EUP

Figure 2. Ages of mitochondrial haplogroups in Europe (from Richards et al., 2000).EUP - Early Upper Palaeolithic; MUP Middle Upper Palaeolithic; LUP Late Upper Palaeolithic.


4/12


8


The opponents of this concept did not missthe opportunity to point out that the deepesttime assessment for an in situ European hap-logroup was paradoxically older than the ageof AMH appearance in Europe (Barbujani, Ber-torelle, 2001). It should be noted that the timeestimates are based largely on the calibrationpoints used (mutation rate). But neverthelessthe ability to present a time estimate was thestrongest among of Richards et al. argumentswhereas the contrary concept was based ex-clusively on the similarity between the geneticpattern and that of the Neolithic spread. Thisenabled one to pinpoint the main logical weakpoint in the "Neolithic" concept, namely, thatthe AMH initial settlement of Europe followedthe same geographical trajectory which was lat-er used by expanding Neolithic farmers. WhenBarbujani and Bertorelle (2001) summed up this

discussion they admitted, that the gradual "outof Anatolia" geographic pattern, as establishedby classical and (later) other markers was cor-rect. Yet this pattern could have originated fromboth, the Palaeolithic and Neolithic migrations,as the both were believed to follow the sameAnatolian route (Fig. 3). The lesson learnt wasthat geographic patterns of genetic variation donot allow distinguish between these scenarios,and one needs non-recombining systems whichare essential for time estimations.

Nearly simultaneously with the seminal

publication summarising the mtDNA data (Rich-ards et al., 2000), two papers on the secondnon-recombining system appeared, summingup the paternal perspective, i.e. Y chromosomal

variations in Europe (Semino et al., 2000; Ross-er et al., 2000). The both papers were based onextensive datasets. Although written in a differ-ent manner they established similar features.

Rosser and colleagues followed a phenom-enological approach, describing patterns of Ychromosomal variation. They found very cleargeographical clines in the distribution of all hap-logroups and statistically calculated that the ge-netic similarity of populations was affected bytheir geographic proximity rather than linguis-tic similarity. In contrast, to that Semino andcolleagues following an interpretative approachconcluded that the observed geographical pat-terns could have been caused by the factors ofsimilar geographic distribution in the Palaeoli-

thic epoch. One may easily note, that in doingso they committed the same logical mistake asthey interpreted geographical pattern "by asso-ciation" with the known event of the same spatialpattern. And indeed, having reanalysed Semi-nos dataset, Chikhi et al (2002) came to theopposite conclusion and interpreted the clinesas having been formed during the Neolithic.At that time, time the estimates for Y chromo-somal haplogroups were much less informativeand reliable than those for mtDNA. The reasonfor this is that it is hard to distinguish on theY chromosome the pre-existing diversity (which

founder population had brought from its home-land) and one accumulated in situ. (Founderanalysis, which was the convenient instrumentfor mtDNA, proved to be too complex to be ap-plied to the Y chromosome).

Since the 1990s, the studies on mitochon-drial DNA and Y chromosome diversity becamedominant in population genetics, which result-ed in a specific "two-system" way of thinking.According to it, the greater part of migratoryevents was allegedly reflected in the both sys-tems. Over the following years numerous stud-

ies were published on Y chromosomal and mtD-NA variation in virtually all European countries.Most of them provided missing pieces for the Eu-ropean genetic puzzle but did refrain from mak-ing oversimplified and/or general conclusions.Those which did could be roughly classified intotwo groups: those describing the overall geneticlandscape (based on the data of the totality ofhaplogroups) and deducing a particular geneticevent from the distribution of particular haplo-group (the haplogroup-driving approach).

Mitochondrial landscape of Europe

From the perspective of mitochondrial DNA,the European gene pool consists of 7-10 most

Figure 3. A scheme of the main demographic processes

documented in the archeological record of Europe (fromBarbujani, Bertorelle, 2001).

Numbers are approximate dates, in years before thepresent. Black arrows, Paleolithic colonization; grey

arrows, Late Palaeolithic recolonization from glacial refugia(grey circles); white arrows, Neolithic demic diffusion.


5/12


9


frequent haplogroups. All but one of them camefrom the Near East: in the majority of casesduring the Early Upper Palaeolithic (EUP) inconjunction with the initial AMH dispersal and a

smaller part during the Neolithic epoch (in thecourse of Neolithisation, Richards et al., 2000).The genetic landscape has been reshaped in theMesolithic/Late Palaeolithic times, during therepopulation of Europe from the southern Eu-ropean refugia. The only European haplogroupthat presumably had emerged in Europe (hap-logroup V) became spread across the entirecontinent during Mesolithic/Late Palaeolithic re-colonisation (Torroni et al., 2001). The westernEuropean origin of this haplogroup (the Franco-Cantabrian refugium) is presumed, based on its

high frequency in this area as well as the oc-currence of its phylogenetic predecessor (pre-Vlineages). However, a recent accumulation ofgenetic data on previously poorly studied East-ern Europe, enabled the present author (Bal-anovsky, 2008) to suppose the occurrence ofan additional East European centre of origin ofthis haplogroup. This finding is based on evenhigher frequency and yet again, on the pres-ence of pre-V lineages in East European steppearea. Impossibility to distinguish between thewestern and eastern European homelands em-phasised the important feature of the European

mitochondrial landscape its extreme homoge-neity.

Indeed, when additional data from differentEuropean populations became available the ge-netic similarity in haplogroup frequencies (andidentity in haplogroup spectra) has been im-mediately recognised (Simoni et al., 2000). Asa result, mtDNA studies has appeared dealingwith Europe as a whole, comparing it with theNear East or other areas, while attempts to tracegenetic processes within Europe encounteredproblems (Helgason et al., 2000). This devel-

opment was rather discouraging for archaeolo-gists and linguists who were typically interestedin a genetic support of existing hypothesis intheir respective disciplines, although on a muchsmaller scale. Fortunately, the paper entitled "Insearch of geographic patterns in European mito-chondrial DNA" (Richards et al., 2002) made thepoint that with the emergence of a larger data-set (with more than 3,000 individual mtDNAs) aspatial structuring became more apparent (e.g.the south-north difference was acknowledgedamong macro-regions of Europe: Mediterraneanarea, Central Europe, Scandinavia, and, surpris-ingly, the Basque Country).

Presently one can affirm that size of thedataset is the key factor. Having at our disposal

the database six times larger than previouslypossessed, comprising 20,000 European mtD-NAs (Balanovska, Zaporozhchenko, Pshenich-nov, Balanovsky; MURKA Mitochondrial Data-

base and Integrated Software, unpublished) wewere able to recognise a much clearer patterning(Fig. 4). European populations altogether occu-pying all parts of this plot provide a geometricalillustration of the genetic variation in Europe.

Figure 4. Genetic relationships of European populationsfrom mitochondrial DNA perspective.

A. The multidimensional scaling plot (geometric distancesbetween points display the genetic distances betweencorresponding populations). Ellipses mark populations

belonging to the same linguistic group.

B. The idealized approximation of the plot (A) byflower-like structure. Black core proto-Indo-European

population; dotted ellipses hypothetical extinct linguisticgroups.


6/12


10


The most remarkable feature is that the patterndistinctly resembles linguistic groups: popula-tions cluster together according to their linguis-tic group of the Indo-European family. Three

largest clusters are formed by Romanic, Slavicand Germanic speakers, while Baltic and Celticspeakers form smaller clusters and Albaniansform a "cluster" of its own outside any othercluster. There are a few exceptions: Romanians,Aromuns and Sicilians lie outside the Romaniccluster while Estonians join the Slavic cluster. Inboth cases the geographical distance (remote-ness for Romanians and Aromuns, proximity forEstonians) had probably a stronger impact ongenetics than the linguistic affiliation.

Among the non-Indo-European populations

of Europe the Basques found their place outsideany other cluster but close to the Romance one(not surprisingly, considering the geographicproximity again). Finno-Ugric and Turkic speak-ers are not shown on this plot because of theirextreme genetic variation, but on another plotthey lie apart of Indo-Europeans.

This linguistic structuring of European mi-tochondrial DNA follows a remarkable "flowershape" pattern: all clusters looking like petalsaround the "core" of the flower. The possible ex-planation of this pattern is that genetic and lin-guistic differentiations were parallel processes

or, in better words, two aspects of the sameprocess, related to the multiplication and differ-entiation of proto-Indo-European population inEurope. It is well known, that in many particularcases distribution of genes is opposed to distri-bution of language (especially in cases of thelanguage replacement by the elite dominancemodel). However, in very general view, almostall Europe is populated by speakers of one lin-guistic family and almost all Europe is geneti-cally homogenous. This allows speculations (likeour flower-like interpretation of the genetic plot)

which consider genetic and linguistic evolutionas generally parallel processes, disregardingpartial exceptions. Such speculations inevitablyoversimplify both processes but could serve as astarting point for more detailed studies.

Therefore, one can accept as a workinghypothesis the differentiation of proto-Indo-European language into linguistic groups beingaccompanied by genetic differentiation resultingin a clear clustering pattern (Fig. 4). This allowsone to introduce time frames into the forma-tion of the European mitochondrial landscape. Itwould coincide with origin and differentiation ofEuropean branches of IE family, i.e. covers thelast 5-6 millennia (Starostin et al., their linguis-tic database is avalable at http://starling.rinet.

ru/main.html). This does not necessarily implythe Neolithisation (for example, major changesduring the Bronze Age is one of alternative ex-planations), but lends credence to the hypoth-

eses advocating a relatively recent origin (or atleast late major reshaping) of the mitochondrialpool in Europe.

One may note that the significance of thelinguistic factor is quite obvious on the graph(Fig. 4 A). However, the idea of a single proto-population totally depends on the flower-likestructure of this graph (Fig. 4 B) and should betherefore considered as one of the plausible hy-potheses.

Y chromosomal landscape of the Europe

While the "homogeneity" is the principalfeature of mitochondrial pool, the Y chromo-somal pool is characterized by a high heteroge-neity. As with mtDNA, there are seven Y chro-mosomal haplogroups dominating in Europe.But while frequencies of mitochondrial haplo-groups are quite similar across Europe, Y chro-mosomal haplogroups follow a clear geographi-cal pattern (Fig. 5). Neither classical markers,nor mitochondrial haplogroups demonstratedsuch obvious and elegant trends. Therefore, Ychromosome became an effective instrument in

population genetics.One should remember that European gene

pool cannot be homogeneous and heteroge-neous at the same time. The question is to whatdegree different markers are able to reveal theexisting degree of variations. Having dozens ofautosomal markers, classical population geneti-cists achieved reasonable resolution in assess-ing the variation between populations (Cavalli-Sforza et al., 1994). Mitochondrial DNA failedto reveal a difference between populations andsuccessfully operates only at a higher hierarchi-

cal level: separating regions (Richards et al.,2002) and linguistic groups (present study, fig.4). Y chromosome operates much better andseparates even subpopulations within the sameethnic group (Balanovsky et al., 2008). Recentstudies based on half of million autosomal mark-ers became able to separate even individualswithin subpopulations (Novembre et al., 2008).

This high differentiation power of Y chro-mosome (i.e. clear geographical clines of itshaplogroups) was revealed already in the earlylarge-scale studies (Semino et al., 2000; Rosseret al., 2000). These clines have been recentlysummarized in a panel of frequency distribu-tion maps (Balanovsky et al., 2008). Two mainhaplogroups, accounting altogether almost for a


7/12


11


Figure 5. Geographic distribution of European Y chromosomal haplogroups (modified from

Balanovsky et al., 2008).K number of studied populations; n number of studied individuals; MIN, MEAN, and MAX-minimal, mean and average frequency on the map.


8/12


12


half of the total European Y chromosomal poolare distributed along the west-to-east axis. Hap-logroup R1b accounts for roughly 50% of the Ychromosomal pool in Western Europe and de-

creases eastward, while R1a reaches the samehigh frequency in the east (Fig. 5) and decreas-es westward.

Analysis of another type of Y chromosomalmarkers (microsatellite variation) also provedthe western and eastern domains to be mainfeatures of the Y chromosomal pool (Roewer etal., 2005). As it was stressed above, the "inter-pretation by association" should be made withcaution. That is why attributing these domainsto Late Palaeolithic re-colonisation from twoprincipal refugia (the south-western and south-

eastern ones) can be considered as a possiblebut not yet proven hypothesis. (One of otherpossibly hypotheses is attributing these geneticdomains to descendants of Late Neolithic BellBeaker and Corded Ware cultures).

These two principal European haplogroupsR1b and R1a are shared between Europe andother regions (Central Asia, Near East, Indiaand North Africa). But two other haplogroups,I1 and I2a (according to previously used no-menclature the same haplogroups were labelledas I1a and I1b, respectively) are restricted toEurope, where they had likely originated.

While R1b and R1a occupy the west and theeast, I1 and I2a predominate in the Europesnorth and south, respectively. I1 which is fre-quent in Scandinavia and southern Baltic areahas attracted less attention due to obviously latecolonization of this region. In contrast, the dis-tribution of haplogroup I2a (Balkan haplogroup)has been widely debated. As southeast Euro-pean autochthonous haplogroup it could not beattributed to Neolithic immigrants (or any otherimmigrants) into Europe. We will discuss it inmore details below.

Three remaining haplogroups (E, J, andN1c) are not evenly spread across the entire Eu-rope but are restricted to distinct areas. For thisand other reasons they are believed to marklater migration waves into Europe which did notcover the entire continent.

The haplogroup N1c (N3 or TAT, accord-ing to previous nomenclatures) is restricted tonorth-east Europe (mainly Finnic speakers) andSiberia. During the last decade it remained un-clear whether is marks an eastward migrationfrom Europe or the opposite westward migra-tion trend. In 2007 Rootsi and colleagues haveshown that this haplogroups could be deeplyrooted in East Asian phylogeny and therefore theoccurrence of this haplogroup in Europe may be

attributed to the Asian influence. Authors sup-posed step-by-step migration from North Chinato Eastern Europe, which started in early Ho-locene and underwent a secondary expansion

on its long way. Derenko and colleagues (2007)studied microsatellite variation associated withthis haplogroup in more detail and tried to esti-mate its age. They identified two variants, one ofwhich migrated into Europe 6-10 ky ago, whilethe second (less frequent) variant was shown tocome by the way of a smaller and more recentmigration, 2-4 ky ago. Although these time esti-mations should be taken with great caution, theboth studies (Rootsi t al., 2007; Derenko et al.,2007) agree that north-east Europe had a sig-nificant (or even predominant) genetic legacy in

South Siberian/Central Asian populations.This creates a problem for "two systems"approach, because from mitochondrial perspec-tive Siberian/East Asian haplogroups appearedin low frequencies and only in the eastern edgeof Europe and did not account for a significantportion of the gene pool anywhere else in Eu-rope. (When low frequency of typical East Asianhaplogroup F was found on Croatian isles andin even lower frequencies on Croatian mainland,this was considered as a paradox and a spe-cial paper (Tolk et al., 2001) tried to explain itby possible medieval gene flow caused by trade

routes of Venice). That is why a significant Asianpresence in Europe, concluded from haplogroupN1c remains one of the main inconsistencies be-tween Y chromosomal and mitochondrial genet-ic systems. From our point of view, this problemcould be resolved if one takes into considerationthe fact that genetic boundary between Europeand Asia lies much eastern than Ural Mountains.The western Central Asia (the Altai Mountainsin particular) could be therefore considered asa genetically intermediary in present time andprimary "European" zone in the past. This view

explains why Y chromosomal haplogroup N (de-spite its origin in East Asia 20-30 ky ago) in preNeolithic or Neolithic times could be the char-acteristic haplogroup for Caucasoid populationsin Eurasian steppe west from the Altai and alsofor Mongoloid populations east from the Altai.From the western part of this area the haplo-group N1c could spread northward and north-westward by a number of migrations suggestedfor this area. This view also explains why thesemigrations did not bring East Eurasian mito-chondrial haplogroups into Europe: the sourcearea having mainly Western Eurasian haplo-groups even in the contemporary gene pool. Itwas more the case in earlier times before Turkicspeakers brought East Eurasian haplogroups by


9/12


13


their expansion into this area two millennia agoonward.

Two last Y chromosomal haplogroups tobe discussed are E and J. They predominate in

North Africa and Near East, and in Europe theyare found mainly in Mediterranean area. Not allsub-branches of these two haplogroups reachedEurope, but mainly one branch of haplogroupE (namely, E-V13) and two branches of haplo-group J (J-M241 and J-410).

While J-410 follows the separate pattern,the other two haplogroups (and also haplogroupI2a, mentioned above) are concentrated in theBalkans and have not been found in neighbour-ing regions with any significant frequencies.The ages of these haplogroups estimated from

their STR diversity are: E-V13 from 4 to7,5 ky;J-M241 from 3,5 to 6 ky; I2a from 5,5 and 10 ky(Battaglia et al., 2008). This roughly coincideswith time of Neolithic transition in this part ofEurope. The model suggested by Battaglia andcolleagues states that Neolithic cultural packagewas adopted by local Mesolithic populations ofthe Balkans which started grow in numbers, ex-panding farming across the entire Balkan penin-sula and later transmitting this package to otherMesolithic populations of Europe. This model ex-plains why these three haplogroups are restrict-ed to the Balkans, why they exhibit decreasing

frequency towards other part of Europe and whytheir age is similar to that of the Neolithic transi-tion.

However, the internal logic of this model isopposite to those applied by Richards and col-leagues in relation to mitochondrial DNA. In-deed, Richards and colleagues proved that mi-tochondrial haplogroups whose diversity wasaccumulated in Europe in situ are of Palaeoli-thic age; and from this fact they concluded thatpresent-day Europeans are descendants of Pa-laeolithic population of Europe (Richards et al.,

2000). Eight years later, Battaglia and colleaguesproved that Y chromosomal haplogroups whosediversity was also accumulated in Europe in situare of Neolithic age; but from this contrastingfact they concluded also that present day Euro-peans are descendants of Palaeolithic populationof Europe (Battaglia et al., 2008). Both studiesare reasonably substantiated and their conclu-sions look correct. However this example illus-trates that genetic studies need more robust anduniversal logic, at least when dealing with suchcomplex process like the Neolithisation. In thisparticular case the possible logical compromiselays in the fact that concept of Battaglia and co-authors actually implies both, cultural diffusionand demic diffusion models. Although the au-

thors did not formulate this explicitly, their con-cept implies, that cultural diffusion took placebetween regions (Analolia and Balkans, Balkansand Central Mediterrania) while the demic diffu-

sion occurred within regions.

Ancient DNA

Analysis of ancient DNA (aDNA) providesdirect data on the former European gene pool,which are free from assumptions and specula-tions which often accompany deductions of pastgenetic processes, based on the contemporarygenetic pattern. This advantage of aDNA maytrigger a revolution in population genetics andif this did not happen so far, this was due to

limited quality and quantity of available aDNAevidence.Problems with the quality (authenticity) of

aDNA data are dramatic because of the possiblecontamination by modern DNA. For this reasonsome aDNA results may be false, and manyearly aDNA papers were criticized exactly fromthis point of view. This problem could be par-tially resolved in few high-standard laboratoriesonly, which have special equipment to minimizerisk of contamination. Cross-checking, i.e. in-dependent analysis of the same ancient samplein different aDNA labs is the second condition.

The third one is implementing the "modern DNAfree" style of excavation into the practice of thearchaeological fieldwork. Having these threeconditions met, one can reach reasonably de-gree of authenticity of the aDNA results.

The quantity problem consists in the scar-city and limited sample sizes of the aDNA data.Again, this problem could be solved only par-tially, by increasing the number of aDNA studiesand average sample size per study. Fortunately,both factors tended to increase in the last de-cade, still trace amounts and high fragmenta-

tion of ancient DNA samples hinder its high-throughput analysis.

Because of these limitations aDNA at leastpresently cannot be the main source of geneticknowledge about the Neolithisation. But it isalready one of the important sources on thisproblem. Indeed, analyses of Neandertal mito-chondrial DNA (Krings et al., 1997; Ovchinnikovet al., 2000), though being criticized for prob-able mistakes in sequencing, put an end to alengthy discussion of the possible assimilationof the Neandertal populations by anatomicallymodern humans. Specificity of Neandertal mi-tochondrial type (Currat, Excoffier, 2004) andabsence of this type in present day Europeans(Behar et al., 2007) allow to root European gene


10/12


14


pool in AMH colonization of the Europe, disre-garding the previous epochs.

Direct genetic data on first Neolithic groupsin Europe are of course most promising source

for choosing between the demic and cultural dif-fusion models of Neolithisation. Such data arenow available for Neolithic population of theIberian peninsula (Sampietro et al., 2007) andNeolithic population of the Central Europe (sitesof Linear Band Ceramic with age of 7.0 7.5ky; Haak et al., 2005). Iberian Neolithic popula-tion was shown to be genetically similar to thepresent day Iberian population. In contrast, LBKpopulation in Central Europe was shown to begenetically distinct from the present day popu-lation of that or any other region of Europe).

The most remarkable feature of the Neolithicpopulation was mitochondrial haplogroup N1afound in 6 of 24 individuals. This haplogroupis virtually absent in present-day Europe. Themathematical simulation has shown that if thisNeolithic population was source of present-dayEuropeans they could not have lost this haplo-groups by stochastic genetic drift.

It was therefore concluded (Haak at el,2005) that Neolithic LBK population did not be-come parental for present-day European genepool, but became dissolved in pre-existing Eu-ropean populations. This conclusion is therefore

in agreement with cultural diffusion model in as-suming that since Neolithic farmers arrived inEurope, the farming was adopted by aboriginalpopulations and first farmers did not leave anyconsiderable genetic legacy in their new home-land.

The study by Haak and colleagues did an-swer the question: "what happened with firstfarmers after their arrival in Europe". To addressthe another question, "where these first farm-ers came from", the consequent study was per-formed (Haak, pers. comm.). Based on extend-

ed dataset (44 individual mtDNAs from differentsites of early LBK culture) it was found that thispopulation is genetically similar to present daypopulations of Northern Mesopotamia, southernCaucasus and eastern Anatolia. Although thegenetic composition of this area could be dis-turbed after the Neolithic period by subsequentmigrations, it is reasonable to suppose that in-ner areas of the Near East were homeland formigrating groups who finally brought these mi-tochondrial lineages into the LBK population ofthe Central Europe.

Of course, this data give rise to many newquestions, and currently available aDNA dataare not sufficient to address them. The mod-erate optimism is based on increasing number

and quality of aDNA data which might allow bet-ter chronological and geographical resolution ofgenetic processes in the near future.

Conclusions

The increasingly accumulating genetic dataon extant and extinct (aDNA) European popu-lations are most frequently discussed in termsof two opposite concepts: demic diffusion andcultural diffusion models of Neolithisation. Inhands of Cavalli-Sforza and his colleagues thegenetic mirror reflected Neolithic expansionacross Europe (demic diffusion); but in hands ofpresent-day writers this mirror reflects mainlythe Palaeolithic legacy of Europeans and cultural

diffusion model is needed to explain spread offarming.Understanding the genetic history of Eu-

rope implies clarifying relative significance andpatterns of each of the following processes:

the initial dispersal of AMH in Europe (Upper1.Palaeolithic);the restructuring of the genetic landscape2.during the Mesolithic repopulation of theEurope from two-four refugia;the importance of the Neolithic expansion3.viewed as the spread of early farming com-munities or spread of Neolithic cultural

package;the role of post-Neolithic human movements4.within Europe;the "oriental" influence in different epochs 5.from Palaeolithic to Medieval times.To address these questions population ge-

netics operated with autosomal (classical) mark-ers in the past and autosomal (DNA) markersmay became the new standard in the future,while the present day studies are based on mi-tochondrial DNA and Y chromosomal variation.

Analysis of mtDNA demonstrated that most

of European haplogroups came from the NearEast during the Upper Palaeolithic times andNeolithic migration of Near Eastern farmers didnot contribute much into the European genepool. The south-east northwest cline withinEurope, as established by many genetic mark-ers, is not considered anymore as the trace ofNeolithic expansion, because Palaeolithic colo-nists used likely the same geographical route.

Y chromosomal data reveal distinct do-mains of prehistoric movements within Europe.Particularly, two different haplogroups predomi-nate in Western versus Eastern Europe, and onemay speculate about two secondary homelands,associating them with Mesolithic refugia or cen-tres of later expansions.


11/12


15


Southeast Europe (the Balkans) which de-serves special attention as gates into Europe,is populated by three different Y chromosomalhaplogroups exhibiting similar patterns: being

autochthonous for Europe these haplogroupsstarted to expand in time frames comparablewith the Neolithisation; it was supposed thatthis expansion might took the form of BalkansMesolithic population adopting farming fromtheir Anatolian neighbours.

Analysis of ancient DNA indicated that firstCentral European farmers (LBK) were of NearEastern origin but did not left recognisable de-scendants. The early farmers in Iberia (and pos-sible in other areas of late Neolithisation) wereof aboriginal European genetic type.

Genetic mirror shows a controversial pic-ture: even in this summary "indigenous" Balkanpopulations adopted farming without immigrantfarmers, but "immigrant" gene pool was foundin first farmers even north of Balkans (in CentralEurope). Nevertheless most lines of reasoningshow that Neolithisation did not change drasti-cally the European gene pool and consequentlydid not involve large-scale population move-ments. Since, if one would like to obtain fur-ther information about these (relatively minor)movements from genetic data it is necessary tobe equipped with a large genetic databases and

a good dose of scepticism not to rush to conclu-sions.

References

Ammerman A.J. and L.L Cavalli-Sforza (1984) NeolithicTransition and the Genetics of Populations in Europe,Princeton. N.J.: Princeton University Press.

Balanovskaya E.V. and S.D Nurbaev (1997a) "A computer-aided methods for gene-geographic study of a popula-tion gene pool in space of principal components", Gene-tika (Moscow) Vol 33: 1456-1469.

Balanovskaya E.V. and S.D Nurbaev (1997b) "The selectivestructure of gene pool: possibilities of studying", Gene-

tika (Moscow) Vol. 33: 1572-1588.Balanovsky O, S. Rootsi, A. Pshenichnov, T. Kivisild, M.

Churnosov, I. Evseeva, E. Pocheshkhova, M. Boldyreva,N. Yankovsky, E. Balanovska and R. Villems R (2008)"Two sources of the Russian patrilineal heritage in theirEurasian context" American Journal of Human GeneticsVol. 82(1):236-50.

Balanovsky O. (2008) "Gene pool of the high latitudes inEurasia or where Saami came from?", in Way to North:the enviroments and the earliest inhabitans of Arcticand Subarctic. Moscow, Institut of Geography. P.. 277-282 (in Russian).

Barbujani G, G. Bertorelle and L.Chikhi (1998) "Evidencefor Paleolithic and Neolithic gene flow in Europe", Amer-ican Journal of Human Genetics Vol 62(2):488-92.

Barbujani G. and G. Bertorelle (2001) "Genetics and thepopulation history of Europe", Proceedings of the Na-tional Academy of Sciences USA Vol 98(1):22-5.

Battaglia V, S. Fornarino, N. Al-Zahery, A. Olivieri, M.Pala,M.N. Myres, R.J. King, S.Rootsi, D. Marjanovic,

D.Primorac, R.Hadziselimovic, S.Vidovic, K.Drobnic,N.Durmishi, A.Torroni, A.S. Santachiara-Benerecetti,P.A. Underhill and O. Semino (2008) "Y-chromosom-al evidence of the cultural diffusion of agriculture insoutheast Europe", European Journal of Human Genet-

ics Advance online publication 24 December 2008; doi:10.1038/ejhg.2008.249

Behar D.M, S. Rosset, J. Blue-Smith, O. Balanovsky, S.Tzur,D. Comas, R.J. Mitchell, L. Quintana-Murci, C.Tyler-Smith, R.S. Wells and Genographic Consortium. (2007)"The Genographic Project public participation mitochon-drial DNA database" PLoS Genetics Vol 3(6):e104.

Cavalli-Sforza L.L., P. Menozzi and A. Piazza A. (1994) His-tory and Geography of Human Genes. Princeton: Princ-eton University Press. 1059 p.

Chikhi L, R.A. Nichols, G. Barbujani and M.A. Beaumont(2002) "Y genetic data support the Neolithic demic dif-fusion model", Proceedings of the National Academy ofSciences USA Vol 99(17):11008-13.

Childe V.G. 1928. The Dawn of European Civilization. NewYork, Knopf

Currat M, and L. Excoffier (2004) "Modern humans did notadmix with Neanderthals during their range expansioninto Europe" PLoS Biology Vol 2(12):e421.

Derenko M, B. Malyarchuk, G. Denisova, M. Wozniak, T.Grzybowski, I.Dambueva, and I. Zakharov (2007) "Y-chromosome haplogroup N dispersals from south Siberiato Europe", Journal of Human Genetics. Vol 52(9):763-70.

Haak W, P. Forster, B. Bramanti, S. Matsumura, G. Brandt,M. Tnzer, R. Villems, C. Renfrew, D. Gronenborn, Alt,K.W. and J. Burger (2005) "Ancient DNA from the firstEuropean farmers in 7500-year-old Neolithic sites", Sci-ence Vol 310(5750):1016-8.

Helgason A, S. Sigurethardttir, J. Nicholson, B. Sykes, E.W.Hill, D.G. Bradley, V. Bosnes, J.R. Gulcher, R. Ward, and

K. Stefnsson (2000) "Estimating Scandinavian andGaelic ancestry in the male settlers of Iceland", Ameri-can Journal of Human Genetics Vol 67(3):697-717.

Krings M, A. Stone, R.W. Schmitz, H. Krainitzki, M.Stoneking,and S. Pbo (1997) "Neandertal DNA se-quences and the origin of modern humans", Vol Cell90(1):19-30.

Lewontin R.C., and J. Krakauer J. (1975) "Testing the het-erogeneity of F values" , Genetics. Vol 80: 397-398.

Menozzi P.,A. Piazza A and L.L. Cavalli-Sforza L.L. (1978)"Synthetic maps of human gene frequencies in Europe",Science Vol 201: 786-792.

Novembre J, T. Johnson, K. Bryc, Z. Kutalik, A.R. Boyko, A.Auton, A. Indap, K.S. King, S. Bergmann, M.R. Nelson,M Stephens, and C.D. Bustamante (2008) "Genes mirrorgeography within Europe", Nature Vol 456(7218):98-101.

Ovchinnikov IV, Gtherstrm A, Romanova GP, KharitonovVM, Lidn K, Goodwin W. (2000) "Molecular analysis ofNeanderthal DNA from the northern Caucasus", Nature404(6777):490-3.

Richards M, H. Crte-Real, P. Forster, V. Macaulay, H. Wilkin-son-Herbots, A. Demaine, S. Papiha, R. Hedges, H.J.Bandelt, Sykes B. (1996) "Paleolithic and neolithic lin-eages in the European mitochondrial gene pool", Ameri-can Journal of Human Genetics. 59(1):185-203.

Richards M, Macaulay V, Hickey E, Vega E, B. Sykes, V. Gui-da, C. Rengo, D. Sellitto, F Cruciani, T. Kivisild, R. Vil-lems, M. Thomas, S. Rychkov, O. Rychkov, Y. Rychkov,M. Golge, D. Dimitrov, E. Hill, D. Bradley, V. Romano, F.Cali, G. Vona, A. Demaine, S. Papiha, C. Triantaphyl-

lidis, G. Stefanescu, J. Hatina, M. Belledi, A. Di Rienzo,A. Novelletto, A. Oppenheim, S. Norby, N. Al-Zaheri,S. Santachiara-Benerecetti, R. Scozari, A. Torroni, andH.J. Bandelt (2000) "Tracing European founder lineages


12/12


16


in the Near Eastern mtDNA pool", American Journal ofHuman Genetics Vol 67(5):1251-76.

Richards M, Macaulay V, A. Torroni, and H.J. Bandelt. (2002)"In search of geographical patterns in European mito-chondrial DNA", American Journal of Human Genetics

Vol 71(5):1168-74.Roewer L, P.J. Croucher, S. Willuweit, T.T. Lu, M. Kayser,

R. Lessig, P. de Knijff, M.A. Jobling, C. Tyler-Smith andM. Krawczak M. (2005) "Signature of recent historicalevents in the European Y-chromosomal STR haplotypedistribution", Hum Genet. 116(4):279-91.

Rootsi S, L.A. Zhivotovsky, M. Baldovic, M. Kayser, I.A. Ku-tuev, R.Khusainova, M.A. Bermisheva, Gubina, M. S.A.Fedorova, A.M. Ilume, E.K. Khusnutdinova, M.I. Vo-evoda, L.P. Osipova, M. Stoneking, A.A. Lin, V. Ferak, J.Parik, T. Kivisild, P.A. Underhill, and R. Villems. (2007) "Acounter-clockwise northern route of the Y-chromosomehaplogroup N from Southeast Asia towards Europe", Eu-ropean Journal of Human Genetics Vol 15(2):204-11.

Rosser Z.H., T. M.E. Zerjal, M Hurles,. Adojaan, D. Alavantic,A. Amorim, W. Amos, M. Armenteros, E. Arroyo, G. Bar-bujani, L. Beckman,. J. Beckman, E Bertranpetit, D.G.Bosch, G. Bradley, H.B.Brede, Cooper, P. Crte-Real, R.de Knijff,. Decorte, Y.E. Dubrova, O. Evgrafov, A. Gi-lissen, S. Glisic, M. E.W. Glge, Hill, A Jeziorowska, L.Kalaydjieva, M. Kayser, T. Kivisild, S.A. K. Kravchenko,A. Krumina, V. Kucinskas, J. Lavinha, L.A. Livshits, P.Malaspina, S. Maria, McElreavey, T.A. Meitinger, A.V.Mikelsaar, R.J. Mitchell, K. Nafa, J. Nicholson, S. Nrby,A. Pandya, J. Parik, P.C.Patsalis, L. Pereira, B. Peterlin,G. Pielberg, M.J. Prata, C. Previder, L. Roewer, S. Root-si, D.C. Rubinsztein, J. Saillard, F.R. Santos, G. Stefa-nescu, B.C. Sykes, A. Tolun, R. Villems, C. Tyler-Smith,and M.A. Jobling. (2000) "Y-chromosomal diversity inEurope is clinal and influenced primarily by geography,rather than by language", American Journal of Human

Genetics Vol 67(6):1526-43.Rychkov Yu.G., and E.V Balanovskaya. (1992) "Genetic ge-

ography of USSR population", Genetika (Moscow) Vol .28. (1) : 52-75.

Sampietro M.L., O. Lao, D. Caramelli., M. Lari., R. Pou, M.

Mart, J. Bertranpetit., and C. Lalueza-Fox (2007) "Pa-laeogenetic evidence supports a dual model of Neo-lithic spreading into Europe", Vol Proc Biol Sci. Vol274(1622):2161-2167.

Semino O, G. Passarino, P.J. Oefner, A.A. Lin, S. Arbu-

zova, L.E. Beckman, G. De Benedictis, P. Francalacci,A. Kouvatsi, S. Limborska, M. Marcikiae, A. Mika, B.Mika, D. Primorac, A.S. Santachiara-Benerecetti, L.L.Cavalli-Sforza, and P.A. Underhill (2000) "The geneticlegacy of Paleolithic Homo sapiens sapiens in extant Eu-ropeans: a Y chromosome perspective", Science Vol290(5494):1155-1159.

Simoni L, F. Calafell, D. Pettener, J. Bertranpetit, and G. Bar-bujani (2000) "Geographic patterns of mtDNA diversityin Europe", American Journal of Human Genetics Vol66(1):262-278.

Sokal R.R., and N.L Oden. (1978) "Spatial autocorrelationin biology: I. Methodology", Biological Journal of Lin-nean Society Vol. 10. P. 199-228.

Sokal R.R., N.L Oden, and B.A. Thomson (1999) "A prob-lem with synthetic maps", Human Biology Vol 71 (1):1-13.

Tolk H.V., L. Barac, M. Pericic, I.M. Klaric, B. Janicijevic, H.Campbell, I. Rudan, T. Kivisild, R. Villems, and P. Rudan(2001) "The evidence of mtDNA haplogroup F in a Eu-ropean population and its ethnohistoric implications",European Journal of Human Genetics Vol 9(9):717-23.

Torroni A, H.J. Bandelt, V. Macaulay, M. Richards, .M Cru-ciani, C. Rengo, V. Martinez-Cabrera, R. Villems, T.Kivisild, E. Metspalu, J. Parik, H.V. Tolk, K. Tambets,P Forster, B. Karger, P. Francalacci, P. Rudan, B. Jan-icijevic, O. Rickards, M.L. Savontaus, K. Huoponen, V.Laitinen, S. Koivumki, B. Sykes, E. Hickey, A. Novel-letto, P. Moral, D. Sellitto, A. Coppa, N. Al-Zaheri, A.S.Santachiara-Benerecetti, O. Semino, and R. Scozzari(2001) "A signal, from human mtDNA, of postglacial

recolonization in Europe", American Journal of HumanGenetics Vol 69(4):844-52.

Yamazaki T. and T. Maryama (1973) "Evidence that enzymepolymorphisms are selectively neutral", Nature Vol 245:140-141.

the russian journal of genetic genealogy european prehistory in mirror of genetics a contemporary...

Documents