Размещено на http://allbest.ru

Размещено на http://allbest.ru

The Structure of the Armenian Matrilineal Genetic Pool

INTRODUCTION

Problem statement

The study of historical and demographic events that characterize modern human populations is a primary part of human evolution reconstruction [Garrigan et al., 2006]. The advances in molecular genetics introduced novel approaches into the investigation of the genetic structure of human populations, which alongside with traditional methods might shed light on fundamental issues of worldwide spread and dispersal of Homo Sapiens. The “out-of Africa” concept suggests the south-eastern African origin of anatomically modern humans (AMH) with subsequent eastern and north-eastern expansions [Johanson, 2001; Cameron et al, 2004]. One of the most interesting regions in terms of Homo sapiens's out-of Africa evolution and spread is the Middle East - the “cradle of civilization”, which has played a pivotal role in the history of humankind, witnessing numerous waves of migration of different tribes and peoples at different times. Hence, the study of genetic legacy of ethnic groups, inhabiting today this region, is crucial for the reconstruction of ancient human migratory pathways.

Historical, archaeological, and linguistic evidence prove that since ancient times Armenians have inhabited the territory of the Armenian Highland, which occupies one of crucial areas of the Middle East [Dolukhanov et al., 2004, Arslanov et al., 2007; Pinhasi et al., 2008]. Historical records and numerous studies show that this region was an important ancient migratory pathway for different tribes and ethnic groups, which undoubtedly had an influence on the genetic structure of the Armenian population [Weale et al., 2001; Rootsi et al., 2012; Yunusbayev et al., 2012; Hellenthal et al., 2014]. In the same time, the Armenian population per se is not sufficiently studied genetically since the majority of the projects conducted so far are based only on paternally inherited markers on the Y chromosome [Weale et al., 2001; Yepiskoposyan et al., 2001; Margaryan et al., 2012; Harutyunyan et al., 2009]. The matrilineal component of the Armenian genetic history is still poorly investigated, and only a limited number of publications applied to maternally inherited markers on the mitochondrial DNA (mtDNA) (Nasidze et al., 2004; Harutyunyan et al., 2009; Schonberg et al., 2011).

Aim and Objectives

The main aim of the project is to reconstruct the matrilineal genetic history of Armenians and identify the location of the Armenian population on the genetic landscape of the Middle East using mitochondrial genetic markers, which reflect the maternal component of the population's gene pool.

In order to reach the aim of the research, the following objectives were posed:

1. to assess the distribution of mtDNA lineages in both general Armenian population and in its geographically different groups;

2. to identify region-specific mtDNA lineages in the Armenian gene pool;

3. to assess the level of diversity of Armenian matrilineal genetic structure;

4. to evaluate the rate of genetic relationships between Armenians and neighboring populations of the Middle East;

5. to determine the degree of the genetic affinity between Armenians from different geographic regions;

6. to test the hypothesis of the indigenous nature of the Armenian matrilineal genetic pool.

CHAPTER 1. LITERATURE REVIEW

1.1 The description of the Armenian population based on the genetic studies

Situated at the crossroads of Europe and the Middle East, the Armenian Highland during its long history has served as both a recipient and conduit for gene exchange between the two regions. While archaeological evidence for modern human, as well as the Neanderthal activity in Armenia, during the Palaeolithic exists Arslanov et al., 2007; Pinhasi et al., 2008], the Last Glacial Maximum (LGM) likely made permanent settlements in the region infeasible [Dolukhanov et al., 2004] until the glacial recessions between 16 and 18 kya [Akзar et al., 2007]. Though archaeological evidence for Mesolithic sites in Armenia is sparse [Kartal, 2003], the improving climate during this period allowed the Armenian highlands to gradually transform into a region characterized by abundant water supply and wealth of fertile plains [Redgate, 2000]. These conditions as well as its proximity to the Fertile Crescent catalyzed the region's emergence as one of the earliest agricultural areas (ca. 8 kya) during the Neolithic Revolution [Hovsepyan et al, 2008]. Archaeological investigations suggest a prominent function for Armenia in the trade of culture and ideas during the Neolithic. Furthermore, the Armenian highlands have been correlated with the earliest known developments of leather footwear [Pinhasi et al., 2010] and viticulture [Barnard et al., 2011]; technologies that would later acculturate across the Near East and eventually enter Europe.

Armenians have been living around the area Armenian Highland for thousands of years. The history of Armenians is very long and starts from times that we do not possess many records about. From the beginning the Armenians were not nomadic tribes but rather a sedentary population utilizing the benefits of agriculture. This is important when considering various hypotheses of the origin of Armenians. There is no conclusive opinion regarding the exact origin of Armenians, whether they were the inhabitants of the Armenian Highland from the very beginning or whether they were migrants from a different area, whether their language developed in the area or it was brought by some others who came later and left their language. A number of hypotheses exist regarding the issue of the origins of Armenians, each one of them relying on certain mythological, historical, archeological, anthropological and linguistic evidence; however, not all of the hypotheses have the similar level of strength. However, today the most reliable and trustworthy methodology for defining this or another issue about the population history, is the population genetics.

Several population genetic studies have been performed so far on Armenian population samples. These studies have utilized different genetic systems, such as mitochondrial DNA, Y chromosome, autosomal markers, as well as non-genetic markers such as blood groups and human leukocyte antigen (HLA) system. The majority of the studies, however, considered Armenians as one group in the context of other populations and did not subdivide the sample. This can lead to only rough understanding of the genetic diversity of Armenians and their relationships with other ethnic groups. Since the development and widespread use of modern genetic technologies, population genetics shifted its focus from phenotypic markers to true genetic polymorphisms. Population genetic studies on Armenians have been performed since 2000 both in the wider context of ethnic groups and focusing on local differences within subgroups of Armenians. These studies were conducted predominantly using Y-chromosomal and autosomal markers. The description of several of the most important of them is given below.

In 2001, Weale et al. have analyzed 734 samples of Armenians using 11 single SNPs and 6 microsatellites of Y-chromosome. This is the first study that has provided proper categorization of Armenians living in different geographic areas. Dividing the whole sample into six geographic groups of Armenians yielded an important result: the analysis has revealed a significant geographic stratification of Armenian population and their relative isolation from each other.

In recent years many more SNPs in non recombining portion of Y chromosome have been discovered, thus refining the distinct Y chromosomal lineages. One recent study has genotyped more than 400 samples from four groups of Armenians for 70 binary Y chromosomal markers and 17 microsatellites loci [Herrera et al., 2012]. The most frequent major haplogroups detected in Armenians were R (38% in Ararat valley, 36% in Gardman, 33% in Van, 34% in Sasun) and J (38%, 36%, 43% and 27% in the same groups). Within the haplogroup R, the majority of Y chromosomes belonged to the lineage R1b1b2* (33%, 31%, 32%, 15%, respectively), a predominantly Near Eastern lineage. It is noteworthy that a different lineage of haplogroup R is found commonly in Europe, R1b1b1a. Surprisingly high frequencies of haplogroup T were observed in Sasun (20.1%), while the frequency of this haplogroup is low in other Near Eastern population and possibly originates from Levant. The results of that study suggest Neolithic origin of Armenian population, coinciding with the spread of agriculture in the area [Kushnareva, 1997; Hovsepyan et al., 2008].

Population genetic studies using autosomal genetic markers have also been performed on Armenians. Molecular genetic analysis of Armenians based on HLA markers has been performed recently [Matevosyan et al., 2011] using more than 4000 samples for low resolution typing and 100 samples for high resolution HLA typing. . The most frequent high resolution HLA alleles detected in Armenians were HLA-A*0201 (15.5%); HLA-B*5101 (17.5%), DRB1*1104 (11.5%). In that report Armenians were divided according to geographical regions of Armenia, Karabakh and Diasporan communities. Authors noted general homogeneity of Armenian population, although some structuring was present, such as proximity of Karabakh and Syunik.

Yet, the Armenian mtDNA genetic structure, which directly reflects the maternal portion the gene pool, is studied relatively poor.

To summarize all population genetic studies so far performed on samples of Armenian ancestry, it is necessary to point out one major problem that mostly they have considered Armenians as a single group. Only a couple of studies [Weale et al., 2001; Herrera et al., 2011] have subdivided Armenian samples into several geographic groups and the results of those studies have conclusively reinforced the previous point that it is a mistake to take Armenians as one group [Wells et al., 2001]. Another issue that has repeatedly come up in those studies is the small number of samples used which can lead to biases. And the way those studies have been performed, each with different set of markers, makes it very difficult to perform meta-analysis in order to overcome the problem with small sample sets. In the last 10-15 years there have been a number of articles about population genetics of Armenians, but there are still quite a few unanswered questions concerning the ethnogenesis, history of Armenians and their relations with neighboring populations. Only with large-scale, well-structured projects on sufficiently diverse Armenian samples with comparative data sets of similar quality it will be possible to answer those questions.

1.2 Mitochondrial DNA as a versatile tool for ethnogenomic studies

The applications of molecular genetic markers in population genetics

The progress of molecular biology in past two decades is characterized by the extensive improvement of nucleic acid extraction, purification and especially sequencing methods. These advances provided novel molecular genetics tools for evolutionary, ecological and forensic studies, which allow resolving problems of this disciplines in more accurate way. Rapidly accumulating data on human genetic variation, collected on different types of DNA markers, have been widely used for the reconstruction of genetic history of modern humans [Cavalli-Sfоrza et al., 2003]. Today, there are three types of molecular genetic markers used in population genetics: autosomal markers, inherited biparently, maternally inherited mitochondrial DNA (mtDNA) genome and the paternally inherited non-recombining portion of the Y chromosome. Among them, mtDNA markers are considered invaluable for human evolutionary and population genetics studies [Cavalli-Sfоrza et al., 2003]. Analysis of mtDNA variation has matured in the course of the past 20 years and has become a versatile tool in the study of our species for the time horizon of the last 100,000 years as well as our relationship to other species [Bandelt et al., 2006].

mtDNA structure and functions

Mitochondria are mammalian cellular organelles that have the function of the oxidative phosphorilation and the formation of ATP. Two distinct genetic systems encode mitochondrial proteins: mitochondrial and nuclear DNA. mtDNA is a small 16 569 base pairs (b.p.) circle of double-stranded DNA which encodes 13 essential components of the respiratory electron transport chain, 2 ribosomal (12S and 16S) and 22 transfer RNA's (Fig. 1).

Figure 1. The structure of human mtDNA. Genes are indicated in orange, rRNA's - in yellow, tRNA's - in green, non-coding regions - in violet. LSP - light strand promoter, HSP - heavy strand promoter, O_L - origin of light strand, O_H - origin of heavy strand.

The mtDNA displacement loop (D-loop, or control region) is a 1.1-kb non- coding region which is involved in the regulation of transcription and replication of the molecule. The D-loop extends from position 16024 to position 576 of the mtDNA and is the largest region not directly involved in the synthesis of respiratory chain polypeptides. The D-loop contains three short regions which, in comparison to the rest of the genome, have a highly variable sequence at the population level: hypervariable sequence (HVS) HVS-I, HVS-II, and HVS-III, corresponding to HVR1, HVR2, and HVR3 in some sources [Brandstдtter et al., 2004a]. The precise definition of the different hypervariable sequences does vary from context to context. The forensic community traditionally took 16024-16365 to be HVS-I, 73-340 to be HVS-II, and 438-576 to be HVS-III [Brandstдtter et al., 2004a]. By contrast, more recent population genetics studies [Brandstдtter et al., 2004b] take wider ranges, particularly in HVS-I in order to capture the phylogenetically important positions 16390, 16391, and 16399 (HVS-I 16024-16400, HVS-II 44-340, and HVS-III 438-576). The first complete sequence of human mitochondrial DNA was published in 1981 [Anderson et al. 1981]. Early evolutionary studies focused on polymorphic restriction sites, but with the rapid development of semi-automated complete mitochondrial genome sequencing, recent work has incorporated full-genome analysis.

Traditionally the human mtDNA is numbered with reference to the light strand according to the original so-called Cambridge reference sequence - CRS [Anderson et al., 1981] of mtDNA. In 1999, CRS was resequenced [Andrews et al., 1999] which has revealed 10 substitution errors, and the revised Cambridge reference sequence (rCRS) was considered as the etalon, so that the other obtained mtDNA sequences are to be compared with it.

Mitochondrial genome as a tool for population genetic studies

The phylogenetic analysis of human mitochondrial DNA is based upon the comparison of mtDNA sequences within and between different human populations. Mitochondrial genome has several unique features, which make it an appropriate genetic system for population genetic studies. Mammalian mtDNA is transmitted as a haploid molecule from mother with low frequency of heteroplasmy [Hauswirth et al., 1986], thus the maternal mtDNA lineage of a particular organism doesn't change, excluding de novo accumulating mutations. The mutation rate of the mtDNA is 5 to 10 times faster than that of the nuclear genome [Brown et al., 1979], reaching 1.64x10^-7 substitution/nucleotide/year for HVRI [Soares et al., 2009] mainly because mitochondria do not have repair enzymes neither for errors occurring during the replication, nor for the damages of the DNA [Cann, 1983]. A specific set of mtDNA mutations defines the mtDNA haplogroups. The last build of global phylogenetic tree of mtDNA variation [van Oven & Kayser, 2009] includes ca. 30 major haplogroups assigned by letters from A to Z with numerous sub-clades (Fig. 2).

Figure 2. The schematic phylogenetic tree of major mtDNA haplogroups. The numbers on the tree nodes indicates mutations, defining the haplogroups. (Behar et al., 2007).

According to their frequency distribution, mtDNA haplogroups are considered to be specific for the particular geographic region, which however doesn't reflect the area of origin of the haplogoup. For instance, haplogroups H, J and T considered as West European, C, D and G - Central Asian, etc.

When researchers began to work on human mtDNA in the 1980s after the publication of the famous Cambridge reference sequence of Anderson and colleagues they first employed only a few restriction enzymes, using them to estimate very simple trees. The genealogical resolution of these trees was so low that they gave quite misleading results, which supported either an `out-of-Asia' origin for modern humans, or even the multiregional model [Bandelt et al., 2006]. After the application of higher-resolution restriction analysis approaches on human mtDNA (Cann et al., 1987), first implemented at the A. Wilson's (UC Berkeley, USA) group, mitochondrial genome became both famous and notorious when used as evidence supporting the `out-of-Africa' model. Finally, in 2000th, the development of fully automated sequencing technologies allowed the massive sequencing of whole mtDNA genomes, which has increased the resolution of mtDNA phylogeny many-folds, opening a new phase of mtDNA research.

1.3 Previous studies on Armenian matrilineal genetic structure

In parallel with the development of DNA sequencing techniques, mitochondrial genome became widely used tool for population genetic research. This matrilineally transmitted genetic system has been also used to study populations of the Near East and Caucasus, including Armenians. However, none of those studies were focused on investigation of Armenian gene pool in particular.

In one of such study a hypervariable region I (HVRI) of mtDNA had been sequenced in 353 samples from 9 Caucasian populations [Nasidze et al., 2001b]. Within this sample set 42 Armenians were included. The analysis showed higher genetic diversity in Caucasian populations compared to Europeans but lower than in the Middle East. According to the article, Caucasian populations clustered together while Europeans formed separate group. The authors concluded that both the Armenian and Azerbaijani population in their history underwent a language replacement process [Nasidze et al., 2001b]. Interestingly, the average genetic distance of Armenians was even slightly smaller to other Indo-European populations than to Caucasians. And it is important to point out that Iranians, a major Indo-European speaking group of the region, was not included in the comparative data sets. Azerbaijanis, on the contrary, were much closely related to other Caucasian population than to Turkic-speaking groups, thus enabling to conclude the occurrence of language replacement in that case.

Nasidze et al. [2004] combined the previously reported Y-chromosomal and mtDNA data [Nasidze et al., 2001b] and added more published data on the populations from the Caucasus and the Near East. However, no additional Armenian samples were included. The additional populations allowed more detailed analysis of genetic relationships between the populations and displayed a geographic rather than linguistic influence on the genetic structure of the populations.

More recent paper of Schonberg et al. [2011] was focused on the investigation of mtDNA gene pool of the Caucasian and West Asian populations, using high-throughput sequencing data of mtDNA genome, which included 147 individuals, representing Georgian (n=28), Azerbaijani (n=30), Iranian (n=30), Turkish (n=29) and Armenian (n=30) ethnic groups. This study has shown a high level of diversity of the populations studied, exceeding that within all of Europe and only slightly lower than the West Asian mtDNA diversity, which might indicate an old age of human populations from this region. The Georgian, Armenian and Iranian all individuals had different sequences, while the Azeri two individuals shared the same sequence, and 27 haplotypes were identified among 29 individuals of Turkish origin. Complete mtDNA genome sequences indicate that the three South Caucasus groups are genetically similar, although they represent three different language families: Indo European-speaking Armenians and Turkic-speaking Azeri cluster with the Caucasian speaking Georgians. However, this work had some limitations. First, the Armenian, Azerbaijanis, Turkish and Iranian samples were collected from the capitals of these countries, without clarification of deep ancestral information of these individuals. Second, the size of datasets was quite restricted, and didn't exceed 30 individuals for each population, which hardly could describe the overall genetic diversity of the whole ethnic group. The data on Armenians was presented by 30 individuals from Yerevan; the authors didn't mention the origin of those people even on the grandparental level, which is important when analyzing Armenians, since the population of Yerevan represent various distinct geographical regions of Armenian Highland, which usually doesn't match to the geographic location of the Republic of Armenia.

The Armenian full mtDNA sequences presented in the paper of Schonberg et al. (2011) were used in the recent paper of Derenko et al. [2013]. Though, 10 mtDNA sequences from the individuals of Armenian descent were added in this work. This study indicates that all populations from the Caucasus region (Armenians, Azeris, and Georgians) are quite dispersed (Fig. 3) though their genetic proximity has been demonstrated by Schцnberg et al. (2011).

Figure 3. Multidimensional scaling (MDS) plot based on Fst statistics calculated from complete mtDNA sequences for population samples from Iran, Anatolia, Caucasus, and Europe. The Populations from Schonberg et al. 2011 labeled with “S” after underscore. The figure is taken from Derenko et al., 2013.

Nevertheless, these 10 additional samples were obtained from the territory of Central Iran and regional residence of their matrilineal ancestors wasn't determined.

Another paper, which included the Armenian mtDNA data as of a representative population of the South Caucasus, was published by Yunusbayev et al. [2012]. Here, the authors have conducted the analysis of uniparental Y-chromosome, mtDNA and autosomal markers. The mtDNA samples of the Armenians (n=37), as well as samples of other ethnic groups was genotyped for HVSI, HVSII and coding region using restriction fragment length polymorphism analysis (RFLP). In this research, the principal coordinate analysis based on mtDNA variation of studied groups placed the Armenians closer to Caucasian ethnic groups than to the populations of the Middle East. The Caucasian populations did not differ in common mtDNA haplogroup frequencies to an extent that would allow the discrimination of geographic subregions or language groups, indicating the proximity of matrilineal genetic component of the Caucasian groups. Additionally, the ADMIXTURE analysis, based on autosomal SNP's data, has shown that the gene pool of the Armenian population was predominantly composed of three components - Middle Eastern, Caucasian and South Asian, however one should consider that the Middle Eastern and Caucasian components were not clearly differentiated in this analysis in general. Considering that the ADMIXTURE approach was applied to the autosomal data, which doesn't include mtDNA markers, it also was of our interest to determine if the matrilineal genetic components of Armenians are the same as for autosomes, which was not carried out in work of Yunusbayev et al [2012]. On the other hand, as soon as this work wasn't focused particularly on the Armenians, especially on the maternal genetic component, the same shortcomings that present in other papers described above concerning Armenian datasets are also present here.

Another quite interesting study, where the Armenian population was included, is the work of Topf et al. [2006], who used ancient and modern mtDNA HVRI sequences of different North European and two Middle Eastern (Armenians and Palestinians) populations for tracing the phylogeographic patterns of population of Britain. Armenians were presented by quite large dataset (n=191), however, as in other papers described above, no any demographic and regional information concerning these individuals were mentioned.

All results mentioned above show that Armenian matrilineal genetic portion was described just in the context of studies about other Middle Eastern and Caucasian populations. All studies that were carried out with the inclusion of Armenians didn't contain the appropriate representative dataset: the number of individuals did not exceed 50, except the paper of Topf et al. [2006] and the origin of those individuals was not clearly determined. These data do not allow making any precise conclusions and inferences on the Armenian matrilineal genetic structure. In summary, it points out that the matrilineal genetic history of the Armenian population is still poorly investigated. Considering the matter that since ancient times and up till now the Armenians occupy the territory of the Armenian Highland, which served as the key migration crossroad for numerous populations, the investigation of the Armenian mtDNA gene pool is crucial not only for the reconstruction of the Armenian history, but also for understanding the genetic component of these ancient migrations, which undoubtedly has affected the Armenian genetic structure. Hence, the current state of the study of the Armenian maternal genetic pool is insufficient for clear understanding of these questions and for the place of the Armenians on the genetic landscape of the Middle East.

CHAPTER 2. MATERIALS AND METHODS

genetic armenian population

2.1 Sample collection and genotyping procedures

In our study, the mtDNA HVSI data of 400 unrelated individuals, whose ancestors inhabited different regions of Armenian Highland (Fig. 4) at least on the grandparental level, were used.

Figure 4. The geographic map of Armenian Highland (circled red)

The blood samples were collected in 2009-2010from different areas of Republic of Armenia. Additionally, we have used the unpublished data of 200 Armenians representing the historical region of Salmast and Khoy, currently located in the northwest of Iran. The samples of buccal swabs were collected in several villages of Ararat region in 2010. Informed consent and information about the birthplace, parents and grandparents was obtained from all the donors. The dataset of 400 individuals was considered as a general Armenian population. Further this dataset was divided to three geographic groups - according to the place of origin of grandmother from maternal line: central Armenians (n=85), eastern Armenians (n=61) and western Armenians (n=43) (Fig. 5). For the inter-population analysis we also included the data from Salmast and Khoy.

Figure 5. Geographic groups of Armenians studied.

Thus, in our work we totally used the unpublished data of 600 individuals, representing both general Armenian population and Armenians from several geographic regions within the Armenian Highland.

Blood samples (4 ml) were collected in ethylenediaminetetraacetic acid (EDTA) containing vacutainers. Buccal swabs were stored in a DNA preservative solution consisting of 0.5% sodium dodecyl sulphate and 0.05 M EDTA for transport purposes.

DNA was isolated by standard phenol-chloroform DNA extraction method. All DNA samples were genotyped for mtDNA HVRI (16024-16400 b.p.) by restriction fragment length polymorphism analysis (RFLP) using standard mtDNA RFLP protocols [Macaulay et al, 1999], with generation of mtDNA haplotypes. 400 samples representing general Armenian population were genotyped in 2010 at the group of Molecular Anthropology, Tartu Biocentre, Estonia. 200 samples from Salmast and Khoy regions were typed in 2010 at the Hammer laboratory, Arizona, USA. Haplogroup assignment of all samples was made using the most recent nomenclature of global mtDNA variations, according to Phylotree build15 [van Oven et al., 2009].

Comparative datasets of different populations were taken from previously published papers and open-access databases (Table 1).

Table 1. Comparative datasets used in our study.

2.2 Statistical and bioinformatical analysis

The frequency analysis of haplogroup distribution of studied populations was made via MS Excel. The data of mtDNA haplogoup frequencies was further used for the construction of distance matrixes based on different genetic distances (GD).

Pairwise GDs F_ST, given by [1]

[1],

where is the variance in the frequency of alleles in different subpopulations and is the variance of allele frequencies in the total population, were estimated based on the analysis of molecular variance [AMOVA] ЦST values, using Arlequin v.3.5 software [Excoffier et el, 2010]. For measuring the statistical significance of GDs between studied populations we used Exact test [Raymond et al. 1995] of population differentiation with 1000 steps of permutations. The values p<0,05 were considered as statistically significant. Other GDs (Nei's D and Reinolds's D²) were calculated by PHYLIP package and were given by formulas 2 and 3, respectively:

[2]

[3],

where m is summed over loci, i over alleles at the m-th locus, and where p₁m_i is the frequency of the i-th allele at the m-th locus in population 1 and 2 [Felsenstein, 1989].

Gene-E software was used for the visualization of obtained distance matrices [http://www.broadinstitute.org/cancer/software/GENE-E].

Principal Coordinate Analysis (PCA) was performed on similarity matrices calculated as one minus genetic distance using Genstat software [Trust, 1995]. The variables used for analysis of general Armenian population were the frequency values of 20 major mtDNA haplogroups (A, B, C, D, F, G, H, HV, I, J, K, L, M, N, R, T, U, V, W, X). Values along the main diagonal, representing the similarity of each population sample to itself, were calculated from the estimated GD between two copies of the same population.

For the reconstruction of the phylogenetic tree of the studied population we used Unweighted Pair Group Method with Arithmetic Mean (UPGMA) algorithm, assuming the constant evolutionary rate for all mtDNA HVRI sequences, with the aid of PHYLIP package.

To assess the summary statistics and gene diversity parameters of the samples, we have converted mtDNA haplotypes obtained by RFLP into mtDNA HVRI sequences from 16024-16400 b.p. positions by Haplosearch - the online tool for mitochondrial haplotype-sequence two-ways transformation using population genetics nomenclature [Fregel, 2011].

The mtDNA HVRI sequences were aligned with rCRS by multiple sequence alignment tool MUSCLE [Edgar, 2005] and further were used to calculate basic parameters of genetic diversity of studied populations using the DnaSP v. 5 software package [Librado, 2009]. These parameters were: number of haplotypes, number of polymorphic sites, haplotype diversity (HD), nucleotide diversity (Pi), average number of nucleotide differences (k) and Tajima's D value. Haplotype (genetic) diversity, one of the main parameters of population heterogeneity, was calculated as described in Nei [1987] [4],

 [4],

where n - the number of samples, k - the number of haplotypes, p_i - the relative frequency of the i-th haplotype.

The Network 4.6.1.0 [Bandelt et al., 1999] was used for median-joining networks construction (http://www.fluxus-engineering.com) of mtDNA HVRI sequences of different populations. The weights for all the loci were assigned by default values of 10. Tо reduce the complexity of the network the reduced median procedure was followed by a median joining analysis. The star contraction and maximum parsimony options were used for further simplification of the networks

The maps displaying spatial distribution of different haplogroup frequencies and genetic diversity were constructed by mapping Surfer v. 11 [Golden Software] package by the gridding method. Geographic coordinates of latitudes and longitude for all populations were based on the sampling centers.

CHAPTER 3. RESULTS AND DISCUSSION

3.1 MtDNA gene pool structure of general Armenian population

To elucidate the composition of mtDNA genetic pool of Armenians, we first performed the frequency analysis of mtDNA haplogroup distribution in general Armenian population and compared the results with those of 28 populations, representing several large geographic regions - North Africa, Middle East, Levant, North and South Caucasus, Europe and East Asia. The mtDNA haplogroup frequency values of Armenians and comparative datasets are displayed in Table 2.

The frequency analysis has shown that the majority (69%) of Armenian matrilineal gene pool is composed of four haplogroups - H, U, J and T, reaching 24%, 18%, 15% and 12% in Armenian population, respectively (Fig. 6). According to their geographic distribution, these lineages are considered as European haplogroups as they are found at very high proportions in the West European ethnic groups. However, the region of their origin is controversial and still actively debated.

Figure 6. The distribution of mtDNA haplogroups in general Armenian population.

Table 2. The mtDNA haplogroup composition of all populations studied.

The Armenian modal haplogroup H, which supposed to be originated in West Asia ca. 20-25 kya [Forster, 2004], is the most common lineage in Europe [Achilli et al., 2004], reaching 55-60% of frequency in western European populations with a decreasing south-eastward cline (Fig. 7). Thus the distribution of this haplogroup in Armenian population is consistent with their geographic location.

Haplogroup U, the second modal haplogroup in Armenians (18%), is found at ca. 11-13% in western European groups [Helgason et al., 2001], and considered as one of the oldest (43-65 kya) matrilineal lineages of this region [Soares et al., 2009].

MtDNA haplogroup J composes 15% of Armenian matrilineal gene pool. This haplogroup originated ca. 32 kya ago [Soares et al., 2009], and today reaches 10-14 % in the western European populations [Helgason et al, 2001].

T haplogroup, originated ~26 kya ago, and currently found at 6-10% in native Europeans, presents 12% of Armenian matrilineal gene pool.

Haplogoups A, B, C, D, G, N, M and F, which are found in high proportion in the population of central and eastern Asia, in summary do not exceed 6% in Armenians.

The summary distribution of these haplogroups (excluding N and M, as soon as only some of their sub-clades are considered as Asian haplogroups) in different populations is demonstrated in Figure 8. Worth mentioning, that these haplogroups are found either in trace amounts (0.5% for F and G) or completely absent (A, B, C, D) in the Armenian mtDNA gene pool. The absence of Asian genetic traits in Arm...