Read any historical or archeological description of Central Asia composed in the 20th century west of Urals, and the first page will tell you in your face that that area belonged to Indo-Europeans, Indo-Iranians, Iranians, Arians and the like, before the invasion of the Türks. Since every child knows that the Iranians as a mass are as blond as the Swedes, the blindingly blond Iranian inhabitants of the Central Asia are assured a first runner up place in the race for a place in history from eternity up to the 5th c. AD. For a time, the second runner-up were the Celts, but because they were promoted by scientists, they were disqualified by the very same science. Which left only the Iranians as the viable solution. The Iranian postulate did not need no scientific scrutiny, like any other widely accepted and canonized postulate. The superb artistry and sophistication of the people could only be Iranian, or as a bare minimum learnt from the Iranians. In other words, Indo-Europeans.

Even though the authors aim to link the mtDNA sequences with the Indo-European genes, and with the non-Iranian, so called Tocharian language, and thus with the Indo-European stock, the study demonstrates a dilution of the Asian genes with Far Eastern genes, a fact well documented in the Chinese annals as Chinese contribution to the Hunnic genes in the form of brides, slaves, prisoners of war, and escapees from the Chinese principalities and statelets. Other documented Far Eastern contributions are the mutual incorporations between Huns, Central Asians, Tunguses and Mongols that took place in the past 2000+ years, and associated cross-flow of genes. The implied expanded interpretation of the Western Eurasian genes as exclusively European genes seems to exceed somewhat the scope of this scholarly research.

This study helps to clarify the debate on the Western and Eastern genetic influences in Central Asia. Thirty-six skeletal remains from Kazakhstan (Central Asia), excavated from different sites dating between the fifteenth century BC to the fifth century AD, have been analyzed for the hypervariable control region (HVR-I) and haplogroup diagnostic single nucleotide polymorphisms (SNPs) of the mitochondrial DNA genome. Standard authentication criteria for ancient DNA studies, including multiple extractions, cloning of PCR products and independent replication, have been followed. The distribution of east and west Eurasian lineages through time in the region is concordant with the available archaeological information: prior to the thirteenth-seventh century BC, all Kazakh samples belong to European lineages; while later an arrival of east Eurasian sequences that coexisted with the previous west Eurasian genetic substratum can be detected. The presence of an ancient genetic substratum of European origin in West Asia may be related to the discovery of ancient mummies with European features in Xinjiang and to the existence of an extinct Indo-European language, Tocharian. This study demonstrates the usefulness of the ancient DNA in unravelling complex patterns of past human migrations so as to help decipher the origin of present-day admixed populations.

1. INTRODUCTION

Recently, hundreds of mummified humans have been discovered in the Tarim Bassin area of Xinjiang, west China, which date between 2000 BP and AD 200-300. These mummies are notable for their European features, (e.g. blond or reddish hair, and straight noses), and reveal an ancient Caucasoid substratum that was present far into the east of Asia (Mair 1995). However, Hemphill (1999) and Hemphill & Mallory (2004) have also noted cranial similarities of some Tarim mummies to populations of South Asia. Ancient mtDNA analysis of some of these remains (Francalacci 1995, 1998) has also demonstrated that they contained a European lineage. It has been suggested that the presence of these individuals in this area is related to the existence of an isolated Indo-European language, Tocharian (Hamp 1998), that seems to have affinities with Armenian, Thracian, Phrygian, Hittite and Celtic (Ruhlen 1996) and that survived in this area until the beginning of the second century AD (Ruhlen 1996). Several recent genetic studies have analyzed classical markers of populations of Central Asia, such as blood groups (Fiori et al. 2000), mtDNA (Comas et al. 1998), and Y-chromosome variability (Perez-Lezaun et al 1999; Wells et al. 2001; Zerjal et al. 2002). These studies have all demonstrated that the present-day populations of Central Asia are genetically extremely heterogeneous, with the presence of well-differentiated east Eurasian and west Eurasian lineages in all populations studied. Various ancient DNA studies have also attempted to reconstruct population movements in Asia, and in particular in China and Mongolia (Oota et al. 1999; Wang et al. 2000; Keyser-Tracqui et al. 2003). The results of two of these studies ... demonstrates the opposite - that the ancient population did not have any specific affinity to the European gene pool, and instead is related to modern southern Chinese populations.

In this study, we have determined the mtDNA sequences and haplogroups from several prehistoric samples from Kazakhstan, a core place within the landscape of Central Asia, to help unravel some of the early steppe migrations. The samples span different time periods and cover both the eastern and western regions. To avoid ambiguous conclusions ... we use several different mtDNA coding-region single nucleotide polymorphism (SNP) markers (Richards et al 1996, 2000; Macaulay et al. 1999; Ingman et al. 2000; Kivisild et al. 2002; Herrnstadt et al. 2002) to corroborate the haplogroups suggested by the HVS-I sequences.

2. MATERIAL AND METHODS

Thirty-six teeth retrieved from Kazakh archaeological sites (table 1) and belonging to different individuals were chosen by D.P. on the basis of their macroscopic preservation and subjected to a DNA analysis at the ancient-DNA laboratory in Barcelona....

3. RESULTS

Twenty-nine out of the 36 samples yielded amplifiable DNA. ...

Most of the retrieved sequences (n= 21, 78%) belong to European or west Eurasian mtDNA haplogroups (HV, H, T, I, U and W haplogroups).....

The eastern Eurasian lineages (22% of the sequences) are represented by sequences belonging to haplogroups M, G and A. ....

...Before the seventh century BC there is a conspicuous absence of haplogroups in East Asia; after that period, samples consist of a mixture of east and west Eurasian sequences, with almost half of them belonging to East Asian haplogroups (6 out of 14 individuals; 43%). The first detection of east Eurasian haplogroups varies between geographical areas.

No east Eurasian lineages appear in east Kazakhstan samples that date to the eighth to seventh century BC or earlier.

In south Kazakhstan, the first Asian lineages appear in the seventh to fifth century BC. In central Kazakhstan, east Eurasian lineages are present in samples from the eighth to third century BC, and in the seventh century BC.

In west Kazakhstan, even samples dating to the most recent period examined do not contain any east Eurasian sequences.

4. DISCUSSION

 The ancient Kazakh samples cluster in nodes already detected in the general mtDNA network of Central Asia. Mitochondrial haplotypes among modern Kazakhs have been reported in two separate studies as follows:

45-31.7% containing haplotypes commonly found in east Eurasian;

0-5% containing haplotypes commonly found in India; and

50-63.4% containing haplotypes commonly found in west Eurasia....

The general west-east Eurasian composition of the prehistoric samples in the period after the arrival of east Eurasian sequences (after the seventh century BC) is... quite similar to the values found in the modern Kazakh population:

east Eurasian (42.9%),

west Eurasian (50%) and

Indian (7.1%)....

The only sequence of Indian origin that was observed, belonging to the M4 haplogroup, originates from a site in the south of Kazakhstan. This fact could correspond to an independent, Indo-Iranian genetic infusion into the steppes.

Another sequence belonging to the M4 sub-haplogroup has been found in present-day Kazakhs, although other populations of Central Asia, such as the Uzbeks, Kyrgyz, Tajiks, Turkmen and Uighurs seem to lack this sub-haplogroup. The haplogroup composition differs slightly between modern Kazakhs and prehistoric samples that originate from prior to the seventh century BC, with variations attributable to time structure and differences in sample sizes (table 5).

Haplogroups present in modern Kazakhs, such as B, F, C, Z, D, R, J and Y, were not observed in the prehistoric Kazakhs. By contrast, two haplogroups observed among the ancient samples, W and I, have not yet been found among modern Kazakhs.

The results also indicate that there is an excess of west Eurasian haplogroups in comparison with those currently found (notably haplogroups H and U. However, this may be attributed to the overrepresentation of the earlier temporal period with only west Eurasian haplogroups.

The observed absence of east Eurasian sequences prior to the eighth to seventh century BC suggests an earlier prehistoric expansion of peoples containing west Eurasian sequences into Asia, that probably went further east, into present-day China. This expansion may be related to the discovery of mummies that contain European features and west Eurasian mtDNA sequences in the Tarim basin, China, as well as the relict Indo-European Tocharian.

An intriguing finding is the occurrence in the ancient sample of sequences from Central Asia that are mainly distributed far into the West, such as the Caucasus and central Mediterranean areas. The latter westward movement of Asian nomads, such as that of the Huns (Ancient Chinese Hsiung-nu, "ferocious slave"), estimated to have taken place around the sixth century BC (Ismagulov 1982), corresponds to the period in which the earliest east Eurasian sequences are found among our ancient samples. In addition, the presence of a haplogroup A sequence found in Siberia (Starikovskaya et al. 1998) and a G2 sequence found in Chinese Han (Yao et al. 2002) points to Siberia and Mongolia as a possible source of such migrations

If there was a human substratum of west Eurasian origin far into East Asia in prehistoric times, what happened to it? Analyses of present-day Han mtDNA sequences from different regions in China detect a very residual presence (less than 5%) of European haplotypes in a few regions. These include Qinghai (east to Xinjiang and Tibet) and Yunnan (north of India and Burma), as well as some coastal regions. Moreover, in a sample from the East Turkestan region adjacent to Kazakhstan (modern Chinese Xinjiang, "New Territory"), no European lineages have been found among the 47 individuals sampled so far (Yao et al. 2002). ... Only two west Eurasian lineages found in modern Chinese correspond to Tl and HV haplogroups, are also found among the prehistoric Kazakhs. This suggests that the genetic legacy of the prehistoric European eastward movement was erased by later Asian expansions, and thus had almost no genetic contribution to the present-day Asians. Such extinction may be related to the demographic processes that also lead to the disappearance of the Tocharian language.

Many more samples, sites and populations will need to be studied to clarify the prehistory of Central Asia. Nevertheless, this study demonstrates that ancient DNA can be a useful tool to help interpret contemporary genetic data in admixtured populations, through the reconstruction of past migrations.

This research was supported by the Direction General de Investigation, Ministerio de Ciencia y Tecnologia of Spain (grants BOS2001-0794 and BMC2001-0772), by Departament d'Universitats, Recerca i Societal de la Informacio, Generalitat de Catalunya (grant 2001SGR00285) and a fellowship AP2002-1065 to M.L.S. M.T.P.G. is grateful to the Wellcome Trust for funding his research. The authors are grateful to Orazak Ismagulov and Ainagul Ismagulova (Department of Anthropology, Central State Museum of Republic of Kazakhstan) for providing them with the ancient samples, and to David Comas for his comments on the manuscript.